Making Data Tangible: A Cross-disciplinary Design Space for Data Physicalization

Designing a data physicalization requires a myriad of different considerations. Despite the cross-disciplinary nature of these considerations, research currently lacks a synthesis across the different communities data physicalization sits upon, including their approaches, theories, and even terminologies. To bridge these communities synergistically, we present a design space that describes and analyzes physicalizations according to three facets: context (end-user considerations), structure (the physical structure of the artifact), and interactions (interactions with both the artifact and data). We construct this design space through a systematic review of 47 physicalizations and analyze the interrelationships of key factors when designing a physicalization. This design space cross-pollinates knowledge from relevant HCI communities, providing a cohesive overview of what designers should consider when creating a data physicalization while suggesting new design possibilities. We analyze the design decisions present in current physicalizations, discuss emerging trends, and identify underlying open challenges.



page 4

page 5

page 6

page 7

page 8

page 16

page 17

page 18


Designing Disaggregated Evaluations of AI Systems: Choices, Considerations, and Tradeoffs

Several pieces of work have uncovered performance disparities by conduct...

Sociotechnical Considerations for Accessible Visualization Design

Accessibility–the process of designing for people with disabilities (PWD...

Challenges in Bridging Social Semantics and Formal Semantics on the Web

This paper describes several results of Wimmics, a research lab which na...

Open Player Modeling: Empowering Players through Data Transparency

Data is becoming an important central point for making design decisions ...

Analysis and Modeling of 3D Indoor Scenes

We live in a 3D world, performing activities and interacting with object...

On the Emerging Area of Biocybersecurity and Relevant Considerations

Biocybersecurity is a novel space for the 21st century that meets our in...

Device and System Level Design Considerations for Analog-Non-Volatile-Memory Based Neuromorphic Architectures

This paper gives an overview of recent progress in the brain inspired co...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1. Introduction

Data physicalization—the practice of mapping data to physical form—sits at the crossroads of various domains, including data visualization, tangible user interaction (TUI), and design111We specifically refer to the “design community” as the Design-oriented venues in CHI where researchers focus on broadening interaction design while promoting new aesthetic and sociocultural possibilities (i.e., the DIS conference and the CHI Design subcommittee). We acknowledge that researchers within the broader CHI community are all designers to some degree, but we use the distinction of the “design community” to emphasize the values that this specific group holds within HCI.. Visualization researchers formally define physicalization as “how computer-supported, physical representations of data (i.e., physicalizations) can support cognition, communication, learning, problem solving, and decision making” (Jansen et al., 2015). While physicalization is a nascent research subdiscipline, the fundamental ideas behind physicalization have a rich history that goes beyond visualization research. Examples of physicalizations have existed for centuries (e.g., Sumerian clay tokens (Joosten,Jan, 1996), Ammaslik wooden maps (Dragicevic and Jansen, ), Marshall Islands stick charts (Dragicevic and Jansen, )) and are also integrated into crafting practices (e.g., steganography (Kuchera, 2018), computational textiles (Kurbak, 2018)). As a result of this extensive and diverse history, data physicalization artifacts are diverse in structure, features, and interactions, appearing in many different contexts.

Crafting physicalizations requires design considerations beyond conventional visualization approaches. While all visualizations must consider the expressivity of their designs (e.g., data encoding and data interactions), data physicalizations must also be mindful of structural considerations (e.g., structural stability, material choice) and the context where the physicalization exists (e.g., audience, location). The diversity and interplay of these requirements can make physically representing data difficult. Data physicalization designers currently face an additional burden of having to navigate different relevant intellectual communities’ set of approaches, theories, and terminologies without a cohesive lens. As a result, designers may be unaware of the knowledge and best practices emerging in other communities such as data visualization (i.e., how to effectively encode data), TUI (i.e., how to amplify the capabilities of the human body and physical world for interaction design), or design (i.e., how to identify and leverage a material’s potential). However, bridging diverse knowledge and practice requires understanding each community’s values, biases, and sets of assumptions. To holistically understand the depth and breadth of the design dimensions data physicalization provides, this paper examines physicalizations from the perspective of researchers from the data visualization, TUI, and design communities.

While previous research has cataloged physicalization artifacts based on semiotics (Zhao and Moere, 2008; Vande Moere and Patel, 2010; Offenhuber and Telhan, 2015), fabrication techniques (Djavaherpour et al., 2021), and visualization tasks (Dragicevic et al., 2020), past work often focuses on a single perspective of data physicalization rather than reflecting on the broader intellectual foundations of the communities in HCI that data physicalization sits upon. We instead aim to synthesize knowledge from across these communities to develop a theoretical design space that highlights cross-disciplinary considerations for creating physicalizations. While design spaces can be characterized by their (1) descriptive power, “to describe a significant range of existing interfaces”; (2) evaluative power, “to help assess multiple design alternatives”, and (3) generative power, “to help designers create new designs” (Beaudouin-Lafon, 2004), this design space is more descriptive by nature to focus on the aforementioned aim. We develop this design space by examining key components of physicalization design and development: who, what, where, and how. Who are the intended audiences (e.g., the public, scientists, analysts)? Where are physicalizations used or showcased? What design decisions contribute to a physicalization’s materiality (i.e., presence in the world)? What interactions does a physicalization afford? How do different communities approach physicalization? (And what can we learn from each other?) These questions help elucidate and describe specific factors of physicalization design and their interrelationships to highlight the emerging patterns in physicalization research.

We developed our design space by surveying physicalization research in two popular databases encompassing relevant communities: the ACM Digital Library (DL) and IEEE Xplore. Our survey aims to capture the rapidly emerging practices of physicalization research since Jansen et al.’s formalization of the field in 2015 (Jansen et al., 2015). We look at work that has been published after this date to examine advances since their initial theoretical effort and limit our scope to recent advances. Our survey used a grounded theory-inspired approach to analyze various interdisciplinary design considerations in 47 physicalizations from 18 different venues. We surveyed these physicalizations according to three facets of design—context (the end-user considerations), structure (the physical structure of the object), and interactions (interactions with both the artifact and data)—deriving thirteen dimensions of physicalization design, illustrated in Figure 1. We developed our codes by drawing on the diverse experience of the authors to help characterize and identify key considerations and open challenges in design, computation, interaction, and materiality (Vallgårda and Redström, 2007; Wiberg and Robles, 2010; Vermeulen et al., 2013; Offenhuber, 2020b). Our design space bridges the knowledge from each community: it provides a cohesive overview of what designers should consider when creating a data physicalization and summarizes relevant trade-offs. We close by highlighting promising opportunities for different communities to learn from each other and offering guidance for future work in the field.

Figure 1. We surveyed physicalizations from across HCI to construct a design space spanning three facets—context (end-user considerations), structure (physical structure of the artifact), and interactions (interactions with both the artifact and data)—over 13 dimensions. While dimensions primarily reside in one facet, each design dimension is not siloed to just one facet but rather can influence other facets: the design dimensions are interrelated. See Section 3.2 for more details. The bar on the outer frames shows the distribution of the design dimension within our surveyed corpus. The length of each bar is determined by how many times the design dimension has been used.

Conceptual Framework for Data Physicalization. This conceptual framework picture is a matrix that reflects the coding process. The matrix contains 3 main columns, labelled as: Context, Interactions, and Structure. The Context column contains Task, Audience, Location and Data Source. The Interaction column contains Sense Modality, Interaction Mediator, and Data Interaction. The Structure column contains Data Duration, Data Scalability, Mobility, Material, Encoding Channel, and Embodiment. The bottom of each column has bars of different lengths for each item. Task has long bars for Reflect, Analyze, and short bars for Educate, Express, Collaborate, Enjoy, and a medium long bar for Unspecified. Audience has a long bar for General Public, and short bars for Creators, Researchers, PWD, Children, Family, Other, and a medium bar for Unspecified. Location has medium bars for Home, Workspace, short bars for Body, Public Space, School, Outdoors, and Medium bar for Unspecified. Data Structure has medium bars for General Statistics, Personal, Environmental, Bio and Medical, Fitness and Wellness, and short bars for Geographic, and Image/Video. Sense Modality has really long bars for Vis, Tactile, and short bars for Aural and Taste. Interaction Mediator has long bar for Directed, medium bar for Sensed and GUI, and a short bar for N/A. Data Interaction has medium bars for Explore, Abstract/Elaborate, Collect, Encode, and short bars for Filter, Connect, Reconfigure, Assemble, Select and Annotate. Data Duration has medium bars for Persistent, Ephemeral, and Permanent. Data Scalability has a long bar for One, and a short bar for Many. Mobility has a long bar for Bounded, and a short bar for unbounded. Material has a long bar for Electronic and a short bar for Non-Electronic. Encoding Channel has medium bars for Position, Color, Form, Light, and short bars for Vibrotactile, Volume, and Texture. Embodiment has short bars for Shape-Changing, Swarm Robotics, Light, TUI, Fluidic Display, Embedded Tags, Voxel Prints, Ediblization, Environment, Person and Activity.

Figure 2. PRISMA flow diagram (Page et al., 2021) on how the final corpus was curated (). See Section 3.1 for our inclusion criteria.

Flow diagram showing corpus curation process. Left side has 4 labels, from top to bottom are Identification, Screening, Eligibility, and Categorization. The row of Identification has two boxes, the left box shows the text Records identified through ACM DL database (n = 520), and the right box shows the text Records identified through IEEE Xplore database (n = 107). Both boxes have down pointing arrows to the next box at the Screening row, that’s labelled Records after duplicates removed, non-archival papers (i.e., posters, extended abstracts, workshops) (n = 411). This box has a right pointing arrow to a box that’s labelled Records excluded (n = 216), and a downward arrow to the box in the Eligibility row, labeled Papers assessed for eligibility* (n = 411). This box has a right pointing arrow to a box that shows Records excluded (n = 364). The Paper assessed for eligibility* (N = 411) has two downward pointing arrows to the Categorization row, left solid arrow to a box that is labeled Papers categorized as eligible using inclusion criteria (n = 42), and a dashed downward pointing arrow to the box that’s labelled Paper included from expert augmentation (n = 5)

2. Background & Related Work

To create a design space for physicalization, we draw on existing definitions of the field and explore past characterizations of physicalization from visualization, TUI, and design.

2.1. What is Data Physicalization?

The definition and practice of data physicalization are constantly evolving. Jansen et al. (Jansen et al., 2015) originally defined physicalization as “a physical artifact whose geometry or material properties encode data” (Jansen et al., 2015). This definition implies that only the physical properties of the artifact encode data. However, Hogan and Hornecker (Hogan and Hornecker, 2012) extend this definition by arguing that data in physicalizations can be “encoded in the shape as well as the behavior of the representation.” Using a physicalization made from spin tops as an example, Hogan and Hornecker contend that our interpretation of the data embedded in the spin tops cannot be fully interpreted until the spin top is set in motion. This example illustrates that our understanding of physicalization must go beyond the physical form and encompass a holistic view that considers the object’s materiality (i.e., the quality of presence in the world (Gross et al., 2014; Wiberg, 2018)), a well-studied concept in design research but less common in visualization. Offenhuber (Offenhuber, 2020b)

, in contrast, contends our understanding of data physicalization should be based on broadening our traditional notion of data (i.e., external data sets) rather than simply focusing on the physical artifact. By employing epistemological and ontological perspectives, Offenhuber illustrates how many physicalization examples are not based on external datasets and highlights our implicit and explicit assumptions about data, especially what we consider data to be. These perspectives illustrate how the definition and values behind data physicalization are still in flux. We, however, do not attempt to define data physicalization. Instead, we acknowledge that the definition of physicalization varies among and even within the communities. This variance necessitates cross-disciplinary synthesis to scaffold collaboration and innovation.

Physicalization designers may view and implement physicalizations according to the practices and values of their native intellectual community. As one example, Friske et al. from the design community reflect on how during the process of creating their physicalization “[d]ata became one material among others, no more or less important in design” (Friske et al., 2020)

. With this perspective, Friske et al. discussed how they had to negotiate between how accurately they wanted to encode the data in the physical representation versus embracing the spirit of the material (i.e., yarn). This perspective stands in contrast to the visualization community’s more cognitively-oriented perspective where data encoding is typically prioritized in terms of its effectiveness in helping users carry out their analytical tasks, such as comparing values or estimating correlation

(Card,S. et al., 1999; Munzner,Tamara, 2015). Jansen et al. (Jansen et al., 2015) outlined the need to understand the effectiveness of physical variables (e.g., how do people perceive sizes in a physicalization (Jansen and Hornbæk, 2015)) as part of physicalization’s research agenda. Offenhuber (Offenhuber, 2020b) cautions that framing physicalization as another traditional visualization application may be a “lost opportunity” that unnecessarily restricts the conceptual and aesthetic implications of physicalization. We reference these perspectives throughout the paper in the context of trends in physicalization research to understand how different communities approach data.

Our goal is to create a cohesive, cross-disciplinary survey focused on the design practices and decisions that researchers engage with when creating physicalizations. This survey considers an artifact as a physicalization if Jansen et al.’s or Hogan and Hornecker’s definitions are satisfied. Jansen et al.’s definition encompasses many of the accepted definitions in current research (see Section 3.2 for our inclusion criteria); however, we augment this definition with Hogan and Hornecker’s inclusion (i.e., data can also be encoded through the behavior of the artifact) to increase the diversity of sampled artifacts, including soft wearables  (Howell et al., 2018; Friske et al., 2020), fluidic displays  (Mor et al., 2020; Pon et al., 2017; Quintal et al., 2016), shape-changing displays  (Taher et al., 2017; Everitt and Alexander, 2017; Nakagaki et al., 2020; Boem and Iwata, 2018), and vibrotactile haptic displays (Aslan et al., 2020; Pepping et al., 2020; Woźniak et al., 2020) (Figure 5). The resulting corpus provides an opportunity to reflect on different theoretical and design perspectives across disciplines.

2.2. Physicalization Theory

To develop broad, cross-disciplinary knowledge, we draw on existing theoretical investigations of physicalization. For example, Dragicevic et al. (Dragicevic et al., 2020) outline historical and recent physicalizations organized by Brehmer’s and Munzner’s typology of abstract tasks (Brehmer and Munzner, 2013): to discover, to present, and to enjoy. This survey highlights examples of physicalizations that illustrate each task but does not consider other attributes of physicalizations that affect their use and comprehension, such as materiality. However, this task-centric perspective primarily reflects the core values of the visualization research community (i.e., designing systems that help people gain specific information from data (Sedlmair et al., 2012)). Further, Dragecevic et al.’s survey focuses more on providing a landscape of artifacts as opposed to a systematic analysis of design factors such as material and location.

Other efforts catalog and describe physicalization by focusing on specific design capabilities and epistemological perspectives, such as indexicality (Zhao and Moere, 2008; Offenhuber and Telhan, 2015), autographical interactions between data and representation (Offenhuber, 2020a), embeddedness (Willett et al., 2017), situatedness (Bressa et al., 2022), and multi-sensory capabilities (Hogan and Hornecker, 2012). These works illuminate the importance of a single critical aspect of physicalization design rather than reflecting on the broad intellectual foundations of the field. More practice-based investigations have explored the interdisciplinary interchange between visualization and design-based fields. For example, Hull and Willett (Hull and Willett, 2017) explore how architectural methods can inform how designers can best engage an audience, encode data into physical form, and determine what kind of workflows can best support data physicalization. Hull and Willet’s work highlights the promise of bridging cross-disciplinary methods to benefit physicalization.

While past work illustrates the promise of physicalization from a range of perspectives, we lack design frameworks that accurately and holistically characterize this emerging research space. We take inspiration from Hull and Willet’s work to bridge communities in a way that produces synergistic knowledge of the range of factors people consider in physicalization development, the interrelations of these factors, and key trade-offs. We build on prior theoretical efforts to formulate our preliminary set of codes (see Section 3.2) and expand on these ideas to illustrate how the knowledge from data visualization, TUI, and design can enrich our understanding of data physicalizations. We focus our survey on HCI research papers and exclude practice-based physicalizations from art and commercial tools. This focus allows us to maintain a manageable scope and minimize assumptions about intention, context, and design deliberations and trade-offs by using concrete, explicit evidence from the papers (e.g., descriptions, implementation details, design specifications) to analyze the author(s)’ intentions. In contrast, artistic and public installations often provide minimal descriptions of the intentions and aims behind the works, requiring us to infer design intention and specifications and increasing the potential for errors in our coding process.

3. Survey Methodology

To construct a design space for data physicalizations, we conducted a systematic survey based on a keyword search to assemble a corpus of physicalizations and augmented our corpus with canonical examples. These examples draw from the diverse experience of the authors, who are experts in visualization, TUI, and design. The survey data was then classified through an open coding approach to identify key design space dimensions. This review qualitatively analyzes work over a six-year period (2015–2021,

; Figure 2) across visualization, TUI, and design. Two authors served as coders throughout the process, working in regular consultation with the rest of the research team.

A keyword-based approach (the keywords are explained in Section 3.1) provides two benefits: (i) it imparts a clear cut-off criterion for papers published under the theme and/or reference of data physicalization and (ii) it can lead to more comprehensive results by efficiently sampling a larger range of possible papers. However, different communities may use different terminology, meaning that keyword use may overlook niche terms of art. The result is also subject to the databases’ primary research focus which can fail to encapsulate how communities beyond ACM and IEEE view data physicalization. To address these limitations, we supplement our keyword search with selected case studies drawn from the collective experience of the research team (five total additional studies). This expert augmentation identifies canonical examples based on the authors’ knowledge of the field, providing critical depth to the design space beyond the initial search.

3.1. Dataset Creation

The broad list of communities in which data physicalization research is published prohibited a systematic review of specific conference proceedings. Instead, we selected archival papers using keywords across ACM DL and IEEE Xplore—two of the most popular databases for data physicalization research that encompass a range of venues (ACM CHI, ACM UIST, ACM DIS, IEEE VIS, etc.). The resulting corpus is available in the supplemental materials222See

Our theoretical framework is based on a corpus of 47 physicalizations selected using the following inclusion criteria:

Date: Though examples of physicalizations date prior to 2015, Jansen et al. coined the term “data physicalization” and formalized the field within visualization in 2015. We look at work that has been published after this date to examine advances since their initial theoretical effort and to limit the scope to recent advances.

Keyword Match: We considered papers where the authors used one of the following terms in the title, abstract, text, or author keywords: data physi- (e.g., data physicalization, physicalisation, data physicality), data materiality, tangible- (e.g., tangible data, tangible visualization, tangible visualisation), and physical- (e.g., physicalization, physical visualization). We chose these terms due to their close relationship to data physicalization (e.g., physical visualization is a synonym for data physicalization (Jansen et al., 2013)) and their use in prior literature.

Eligibility: Our initial search retrieved 627 candidate papers: 520 papers from ACM DL and 107 papers from IEEE Xplore. We then screened the candidate papers to remove duplicates333When we encountered the same work by the same authors, we used the latest published work (e.g., Le Goc et al., 2018 (Le Goc et al., 2018)) and discarded the earlier versions from the corpus (e.g., Le Goc et al., 2016 (Le Goc et al., 2016)). and non-archival papers (i.e., posters, extended abstracts, magazine articles, workshop papers). After screening, 411 papers remained. We then removed any papers that prohibited conclusive coding using the following criteria:

  1. The paper introduced a new technique, briefly using one of the keywords but not explaining how the application or design of the technique can be applied to physicalization (e.g., (Braley et al., 2018; Sahoo et al., 2016)).

  2. The artifact created in the paper did not provide data insights but was predominantly used as an interface control (e.g., (Manshaei et al., 2019; Kaninsky et al., 2018; Jofre et al., 2015)). See the discussion of “Data Physicalization and TUI” in Jansen et al. (Jansen et al., 2015) for further discussion.

  3. The paper presented an experiment focused on outlining empirical guidelines (e.g., (Jansen and Hornbæk, 2016; Stusak et al., 2016; López García and Hornecker, 2021)). We prioritized papers that demonstrated a design process through the artifact produced in the paper.

  4. Data was encoded through sonification or olfaction rather than through the physical properties of an object (e.g., (Nath, 2020; Batch et al., 2020)). Sonification (Madhyastha and Reed, 1995; Kaper et al., 1999) and olfaction (Marco and Gutierrez-Galvez, 2012) are considered their own areas of research and outside of our current scope.

  5. Papers that presented multiple artifacts (Lockton et al., 2020; Karyda et al., 2020), but the individual artifacts presented ambiguous codes. To reduce ambiguity, we prioritized papers that allowed insight into the overall design methodology used to develop an artifact.

Using these inclusion criteria, two research team members each examined half of the corpus to remove any papers that did not meet the criteria or needed further discussion. Papers needing discussion were reviewed by the full research team. After deriving the final dataset, we added five additional examples (Boem and Iwata, 2018; Bader et al., 2018; von Ompteda, 2019; Raidou et al., 2020; Nachtigall et al., 2019) based on recommendations from the team, who reflect expertise in visualization, TUI, and design, to capture relevant examples not included in the original corpus.

Design Dimension Code Multiple
Task Analyze, Educate, Express, Reflect, Enjoy, Collaborate, Unspecified Yes
[0.2em] Audience General Public, Children, Family, Researchers, People with Disability, Creator, Other, Unspecified Yes
[0.2em] Location Home, Workspace, Public Space, Body, Outdoors, School, Unspecified No
[0.2em] Data source Biological and Medical, General Statistics, Personal, Fitness & Wellness, Geographical, Environmental, Image/Video Yes
Embodiment Shape-Changing Display, Swarm Robotics, Fluidic Displays, Voxel Prints, Light Display, Embedded Tags, Edibilization, TUI, Environmentally-Embedded, Personally-Embedded, Activity-Embedded No
[0.3em] Material Non-Electronic, Electronic No
[0.2em] Encoding Channel Color, Form, Position, Volume, Texture, Light, Vibrotactile Yes
[0.2em] Mobility Unbounded, Bounded No
[0.2em] Data Scalability One, Many No
[0.2em] Data Duration Ephemeral, Persistent, Permanent No
Interactions Mediator GUI, Direct, Sensed, N/A Yes
[0.2em] Sense Modality Visual, Tactile, Aural, Taste Yes
[0.2em] Data Interactions Select, Explore, Reconfigure, Encode, Abstract/Elaborate, Filter, Connect, Annotate, Collect, Assemble Yes
Table 1. The final codebook with 13 design dimensions and 74 subcodes. Dimensions were further categorized into those describing the context, structure, and interactions of a physicalization. Multiple refers to situations where multiple codes may describe a single dimension according to the paper’s images and supporting text. Full code definitions can be found in the supplemental materials.

3.2. Codebook

We coded each example in our corpus using a grounded theory-inspired approach. Our codebook consisted of terms and categories that examined the three key facets of physicalization design and development (context, structure, and interaction), focusing on five questions: Who are the intended audiences? Where are physicalizations used or showcased? What design decisions contribute to a physicalization’s materiality? What interactions does a physicalization afford? How do different communities approach physicalization? (And what can we learn from each other?) We generated terms and categories that described the variation within our corpus guided by these questions and past theory from relevant communities (e.g., using interaction models from data visualization (Yi et al., 2007) and design (Vermeulen et al., 2013)), adding terms as we observed new design variations that did not fit our initial codes. These codes are intended to reveal the interrelationships between key factors of data physicalization by drawing upon existing theories from the relevant communities. We limited our codes to attributes of the physicalization that could be assessed through the paper’s images and supporting text. This constraint precluded analysis where we would have to make assumptions about the authors’ intention or approach. The final codebook is summarized in Table 1 and full code definitions can be found in the supplemental materials.

We summarize the 13 dimensions of our design space by whether they pertain to a physicalization’s context (i.e., aspects of the design environment such as users, data, or location), its structure (i.e., its physical form), and its interactions (i.e., how it affords and responds to user inputs).

3.2.1. Context

Context describes the physical, social, and analytical frames a physicalization addresses (Chalmers, 2004). Context grounds the visual design and functionality behind a data physicalization, including the attributes describing the intended goals, environment, and users of the physicalization. We use the concept of situated visualization—embedding data into its context of use—from the visualization community as a lens to see how physicalizations make use of context. Context also examines the end-user considerations guiding design, including its intended tasks (what is the functional purpose of the physicalization), audience (who is the physicalization intended for), location (where does the physicalization exist), and data source (what kinds of data are represented).

3.2.2. Structure

Structure focuses on how data is physically embodied and encoded, including its overall form. We draw on theoretical perspectives from the TUI and design communities on embodiment and materiality to characterize structure. Categorizing physicalizations based on the physical structure is challenging, largely due to the diverse fabrication approaches that physicalization designers adopt (Djavaherpour et al., 2021). We focus on how the data is embodied—how data is physically manifested 444Our use of embodiment refers to providing a tangible or visible form to something abstract (Pahl and Beitz, 1996). This definition is distinct from embodiment as considering human bodies and actions in tangible and social computing systems, as described by Dourish (Dourish, 2001).. The concept of embodiment and its benefits have been extensively discussed in the TUI community (Ishii and Ullmer, 1997; Hornecker, 2005; Wyeth, 2008). For example, Zhao & Vande Moere (Zhao and Moere, 2008) who also recognize the importance of embodiment for physicalizations.

To holistically understand embodiment in physicalization, we also examine the physicalization’s use of materials, such as 3D printing filament, yarn, motors, and liquids. Materials are the building blocks of physicalizations, shaping a physicalization’s physical structure, interactions, and materiality (i.e., presence in the world (Gross et al., 2014; Wiberg, 2018)). To understand the relationship between materials and other dimensions of physicalization design, we coded examples of what materials the physicalization is composed of; how data is encoded through the physicalization’s tangible and visual properties; how mobile the physicalization is; whether the physicalization is scalable in handling multiple, distinct datasets; and how long the data is displayed for (i.e., data duration). Through these design dimensions, we can also gain a better understanding of the values designers place in the material. Gross et al. (Gross et al., 2014) state there are three distinct perspectives a designer can consider when using a material: physical, computational, or cultural. The physical framing focuses on how a material functions with its physical properties and how effectively a material can accomplish a task. The computational framing denotes that computers, digital information, and other intangible aspects of computational systems are also considered materials that are incorporated with physical things. The cultural aspect focuses on examining materials related to crafting practices and how materials “communicate” to the maker.

3.2.3. Interactions

Interaction focuses on the dialogue between the user and the physicalization. With physicalizations, a data interaction (e.g., filtering a dataset) can be examined by analyzing how the interaction was executed at the artifact-level (i.e., what design feature in the artifact made it possible to execute the filter interaction?). We draw on the concepts of affordances, feedforward, and feedback (Vermeulen et al., 2013) to examine how interactions in general (e.g., pressing a button) are possible at the artifact-level, and we use Yi et al.’s data interaction framework (Yi et al., 2007) to identify the different data interactions imposed on physicalizations.

We characterize interactions based on how the physicalization affords interactions and what actions are taken with the data. The interactions mediator refers to how an individual’s data interactions are captured (e.g., through a direct action, use of GUI, sensed through a sensor). The sensing modality highlights which human senses are used when receiving feedback. Data interactions examine how each interaction changes the data presented to the user. We use Yi et al.’s framework of data interactions (Yi et al., 2007) to help capture the user’s intent while interacting with the physicalization. We added three additional interactions (annotate, collect, assemble) based on observations from our corpus that incorporate actions that are more natural in physical rather than digital systems. Annotating refers to adding supplemental information to the physicalization, collecting provides additional data to the physicalization, and assembling focuses on piecing physicalizations together.

3.3. Coding Process & Agreement

Two coders independently coded the physicalizations using a preliminary set of design dimensions. They reconciled codes after the first 12 papers to ensure common understanding and application and again after each new code was added. The inter-rater reliability of the preliminary codes using Krippendorff’s alpha (Krippendorff, 2004) was high (; 95% agreement). The coders discussed any conflicting codes with the rest of the research team to arrive at a consensus, with one set of final codes assigned to each physicalization.

4. Design Space

The corpus consists of a total of 47 papers. This corpus covers several conferences in HCI and Design, including CHI (14/47), DIS (10/47), TEI (5/47), UIST (2/47), ASSETS (2/47), NordicCHI (2/47), PerDIS (2/47), AVI (1/47), ICER (1/47), and ICMI (1/47). We also survey conferences from Visualization and Visual Analytics (4/47) 555IEEE VIS (2/47), IEEE VISAP (1/47) and journal articles (3/47).666The Proceedings of the ACM on Human Computer Interaction (HCI), Springer AI & Society, and Science Advances From our coding process, we derived 13 design dimensions that describe critical decisions in physicalization, which we organized into three facets: context, structure, and interactions (Section 4.1 - Section 4.3). We use these three facets to discuss the five questions we posed in Section 1: Who are the intended audiences? Where are physicalizations used or showcased? What design decisions contribute to a physicalization’s materiality? What interactions does a physicalization afford? How do different communities approach physicalization? (And what can we learn from each other?). See Figure 1 and to see how the 13 design dimensions are distributed throughout the corpus.

4.1. Context

Tasks: While the data visualization community typically designs for exploration (Card,S. et al., 1999; Munzner,Tamara, 2015), we found reflection—the act of increasing one’s self-knowledge—to be the most popular task physicalizations support (18/47). Topics of reflection varied with focuses including personal informatics (e.g., exercise (Khot et al., 2016; Sauvé et al., 2020b), reading habits (Ju et al., 2019), emotions (Howell et al., 2018)) and data about family (e.g., children’s emotions (Pepping et al., 2020)). People also used physicalizations as a way to make a personal connection to bigger communities, such as commemorating a local river flood catastrophe (Pon et al., 2017), permit violations in a local river (Perovich et al., 2021), and climate change (von Ompteda, 2019).

The next most common task was analysis (15/47). In these cases, people used physicalizations to explore statistical datasets. Physicalizations were also used to educate (3/47), collaborate (2/47), express (2/47; i.e., using a physicalization to express their feelings or ideas), and enjoy (1/47). There were also a handful of instances (8/47) where the task was unspecified. In these cases, physicalizations were introduced but their intended purpose and/or what users should gain from the artifact was not explicit.

Audience. We found that the majority of physicalizations targeted the general public (21/47), focusing on incorporating data into everyday activities (Nissen and Bowers, 2015; Howell et al., 2018; Boem and Iwata, 2018; Ju et al., 2019) that required little visualization literacy. Beyond the general public, we found physicalizations designed with a specific audience in mind, including creators (6/47) (i.e., the physicalization is meant for the person who created it), researchers (4/47), children (2/47), people with disabilities (2/47), and family (1/47). We found four instances of other (4/47) designed for musicians (Pon et al., 2017), golfers (Woźniak et al., 2020), sports fans (Khot et al., 2016), and college campus visitors (Walker et al., 2017). Several papers (8/47) did not explicitly state the intended audience of the artifact. In these cases, the physicalizations were designed to showcase a novel technology where they did not further explain their intended audience. While many physicalizations are often created with one audience group in mind, Torrent (Pon et al., 2017) (Figure 6a) is an interactive system that augments concert flute music with live water sounds designed for both performers and their audience throughout a performance. Water surface turbulence encodes the flutists’ muscle tension, and the physicalization is situated between the performer and audience to provide information throughout the performance.

Location. Physicalizations in the corpus were designed for a range of locations, such as workspaces, homes, public spaces, outdoors, and even on the body. Many of the physicalizations were meant for a private location, including homes (10/47), the body (5/47), and workspaces (8/47). About a quarter of the papers focused within a public space (8/47), such as schools (Gendreau Chakarov et al., 2020; Veldhuis et al., 2020), golf ranges (Woźniak et al., 2020), and event venues (Pon et al., 2017; Daniel et al., 2019). Two physicalizations specifically focused on natural environments (ocean (von Ompteda, 2019) and river (Perovich et al., 2021)). Approximately a third of the papers in the corpus (14/47) provided little to no information about the physicalization’s intended location. As with the physicalization’s audience, papers with an unspecified location often used physicalization as part of an illustrative example for a new technique.

Physicalizations most commonly resided in homes (10/47). These physicalizations focused on capturing data from the house (e.g., environmental data (Houben et al., 2016), children’s emotions during gaming (Pepping et al., 2020)) or mimicking everyday objects, such as clocks (Lee et al., 2020), lamps (Ju et al., 2019), mementos (Khot et al., 2016), and design pieces (Sauvé et al., 2020b; Menheere et al., 2021; Lee et al., 2021). Physicalizations for the home were designed to seamlessly integrate into the home environment. Everyday objects designed and situated for private environments, such as the home, encouraged more private and personally meaningful reflections (Pepping et al., 2020; Ju et al., 2019; Sauvé et al., 2020b). Laina (Menheere et al., 2021), for example, is a 2.5D shape-changing display designed as an art piece that represented the user’s running route data. Laina’s slow-motion design features allowed the users to reflect on their running path and feel a sense of pride in their accomplishments. Similarly, people who used Bookly (Ju et al., 2019) mentioned that Bookly’s design, which encouraged readers to place and pick up the book at the base of the lamp opened to their current progress, prompted them to read more and be more reflective about their reading habit.

Data Source. Compared to traditional visualization (Sarikaya et al., 2018), few physicalizations focused on traditional static, tabular data (11/47). Target tabular datasets included the U.S. population (Suzuki et al., 2019), world GDP (Taher et al., 2017; Nakagaki et al., 2020), civic data (Claes and Moere, 2015), conference metadata (Tejada et al., 2020), student’s grades (Le Goc et al., 2018), and sports matches (Khot et al., 2016). Several physicalizations supported biomedical and medical data (8/47), primarily using bio-sensing data, such as galvanic-skin response (Howell et al., 2018), gait (Nachtigall et al., 2019), heart rate (Boem and Iwata, 2018; Lo et al., 2016; Khot et al., 2017), respiration rate (Sabinson et al., 2021), and muscle tension (Pon et al., 2017). Fitness and wellness datasets (7/47) also included bio-sensing data, but overall aimed to look at the data for more of a wellness perspective, in terms of improving their physical ability (Woźniak et al., 2021, 2020), setting exercising goals (Khot et al., 2016; Sauvé et al., 2020b), and bringing more awareness of their physical activity (Menheere et al., 2021) and emotions (Howell et al., 2018). Physicalizations also frequently supported reflection on other forms of personal data (10/47), such as one’s reading habit (Ju et al., 2019), schedule (Lee et al., 2020), diet (Sauvé et al., 2020a), fabrication lab practices (Gourlet and Dassé, 2017), and Twitter data (Nissen and Bowers, 2015; Khot et al., 2016). Finally, there are physicalizations that focus on encoding environmental data (10/47), touching upon topics such as energy (Daniel et al., 2019; Quintal et al., 2016; Nakagaki et al., 2015), weather (Suzuki et al., 2018), and climate change (von Ompteda, 2019), and geographic data (3/47). Documented (Ettehadi et al., 2021) was the only artifact that used an image/video dataset (1/47).

4.2. Structure

Shape-changing interfaces were the most popular way physicalizations embodied data (9/47). Other forms of physical embodiment included swarm robotics (5/47), TUIs (4/47), fluidic displays (3/47), light displays (3/47), voxel prints (2/47), embedded tags (1/47), and edibilization (1/47). Data was also physically embodied through environments (9/47), people (6/47), and activities (4/47). See Figure 5 for an overview of physical embodiments. Several physicalizations (13/47) attempted to co-opt existing idioms behind common visualizations (e.g., line chart, bar chart, scatterplot). However, others used a custom data representation, leveraging the unique affordances of physicalization (see Section 5.5.2 for more details).

The majority of surveyed physicalizations incorporated electronics as part of their structure (35/47), while only a quarter of the physicalizations (12/47) did not contain any form of electronics (i.e., non-electronic). Though the majority of the physicalizations incorporated various solid materials (43/47), five physicalizations incorporated liquids as part of their main material choice (Mor et al., 2020; Pon et al., 2017; Quintal et al., 2016; Perovich et al., 2021; von Ompteda, 2019). Liquid displays leveraged a number of physical attributes to encode data, including microfluidics (Mor et al., 2020), use of vortexes as metaphors for energy production (Quintal et al., 2016), surface turbulence in a pool reflecting muscle use during a musical performance (Pon et al., 2017), and as the location for situated data (von Ompteda, 2019; Perovich et al., 2021). We also observed a single instance where the researcher embraced computational materiality (Vallgårda and Redström, 2007)—the perspective that computers, digital information, and other intangible aspects of computational systems are also considered to be materials that can be incorporated with a physical object. Documented (Ettehadi et al., 2021) contains an embossed visual tag that was designed to be embedded within a 3D printed object. To support DIY makers, it focuses on retrieving documentation images and videos using a smart device where the digital information contributes to the 3D object’s materiality.

We observed a connection between the use of materials and whether data is represented ephemerally (i.e., data is displayed for a short period of time), persistently (i.e., data is displayed for as long as the user desires), or permanently (i.e., data is permanently embedded into the physical structure of the physicalization). Approximately half of the physicalizations were able to display data persistently (25/47) through programmatic means, while 12 displayed data ephemerally (12/47) by streaming data from sensors to outputs such as LEDs (Lo et al., 2016; Gendreau Chakarov et al., 2020), vibrotactile motors (Aslan et al., 2020; Pepping et al., 2020), and pneumatics (Sabinson et al., 2021) in real-time. Ten artifacts displayed data permanently with non-electronic materials (10/47). Despite how long data was displayed, position was the most popular data encoding method (23/47), followed by color (17/47), form (15/47), use of light (12/47), vibrotactile (7/47), volume (5/47), and texture (2/47).

4.3. Interactions

The surveyed physicalizations supported 61 total interactions. These interactions were categorized using the seven data interaction categories from Yi et al., (Yi et al., 2007) as well as annotate, collect, assemble (see Section 3.2.3). Of the ten data interactions, explore was the most pervasive (12/61), followed by abstract/elaborate (11/61), collect (9/61), encode (8/61), filter (5/61), reconfigure (5/61), connect (4/61), assemble (3/61), select (2/61), and annotate (2/61). The majority of physicalizations enabled direct interactions (i.e., introducing interactions by requiring users to intentionally touch or manipulate the representation of data; 41/61), while some relied only on sensors (9/61) or required users to interact with a GUI (7/61).

5. Findings

Section 4 summarizes key distributions of different design decisions across the 13 design dimensions. For brevity, we now use those findings to discuss patterns in the design space that help bridge dimensions or disciplines; reveal how structure, context, and interactions are interrelated to each other; and offer compelling future opportunities.

5.1. Who are the intended audiences?

Physicalizations commonly targeted the general public (21/47), focusing on how to incorporate data into everyday activities (Nissen and Bowers, 2015; Howell et al., 2018; Boem and Iwata, 2018; Ju et al., 2019) that required little visualization literacy. Our observation matches Jansen et al. (Jansen et al., 2015)’s description of the benefits of data physicalization—active perception and engagement with the data—as physicalizations commonly target accessible ambient knowledge and awareness of pervasive personal or community issues. However, people’s intended engagement varied throughout our corpus. For example, Physikit (Houben et al., 2016), a system of four cubes with environmental sensors embedded inside each cube, provides active engagement. Each cube outputs the sensor data (e.g., noise, humidity, temperature) either through light, vibration, movement, or airflow. Due to its small size and discrete nature, people actively placed the Physikit cubes throughout their homes to compare environmental sensor data in the different parts of their house. Bookly (Ju et al., 2019), in contrast, focuses on more ambient engagement. Bookly is an interactive lamp that tracks an individual’s accumulated reading time in addition to other reading metrics. Bookly’s design seamlessly integrates reading progress data through a mountain-shaped holder. The mountain-shaped holder provides an intuitive place where users should place their book and emphasizes that the reading is still ongoing by keeping the pages open. Bookly’s overall form enables the system to naturally collect data by tracking a user’s reading time once the book is picked up. Through its volumetric disk, Bookly ambiently displays the user’s accumulated reading time with respect to their reading goal while the pages of the open book communicate progress.

Beyond the general public, several physicalizations were designed with a specific audience in mind, including researchers (4/47), children (2/47), and people with disabilities (2/47). Physicalizations meant for these specific sets of users focused on data exploration tasks. However, the depth of the supported data analysis varied depending on the interactions the physicalizations supported. Zooids (Le Goc et al., 2018) and EMERGE (Taher et al., 2017), for example, provide analytical insight from the rich number of data interactions both systems provides. Zooids is a swarm robotic system where users are able to map data instances (e.g., a student) with multiple attributes (e.g., grades, gender, letter recommendation score) to each robot. The system provides users the ability to encode, abstract/elaborate, annotate, and filter the resulting multi-dimensional data. Based on the combination of these data interactions, Zooids support collaborative decision-making, such as faculty deciding on student admissions or a family planning a vacation. EMERGE (Taher et al., 2017) similarly integrated filter, reconfigure, abstract/elaborate, and select into their 2.5D shape-changing display where users present and share analytical insight from a dataset. Robographics (Guinness et al., 2019) and FluxMarkers (Suzuki et al., 2017) encourage data exploration for users with disabilities, specifically addressing the needs of blind and low vision (BLV) people. However, Robographics and FluxMarkers supported a limited range of interactions. Both systems supported only elaborating data details (e.g., the minimum/maximum value of an x-axis) by combining auditory and physical interactions.

We also found six instances where the physicalization was meant for the creators themselves. Often, these physicalizations are in the domain of personal informatics (i.e., personally relevant information for the purpose of self-reflection and self-monitoring (Elsden et al., 2015)). In the process of creating the physicalizations, creators tended to use custom data mappings to incorporate meaningful reflections (see Section 4 to see the topics of reflections illustrated in the corpus). When analyzing and exploring the physicalization, only the creators would have insights on what the data is about. ListeningCups (Desjardins and Tihanyi, 2019), for example, are tactile cups where each bump and groove on the surface of the cup represents the everyday ambient noise within the authors’ environment. Aside from the authors who created the artifacts, others would not have an immediate context as to what data has been encoded. The authors in contrast described how they gained personal insights from their data when they held the cups during meandering conversations or while washing the dishes. This led to reflections of self-inquiries such as, “Why is the part feeling more smooth, why were we so quiet?”. Friske et al. (Friske et al., 2020) explore tensions between physicalizations use by the creator and by external audiences. They argue that a personal informatics physicalization does not have one “true” data interpretation. They question the authority of data, noting how no one has the authority to invalidate someone’s data interpretation even if they did not create the artifact. They elaborate that these interpretations should not be seen as distinct but as informing each other and the ongoing understanding of the data itself.

Figure 3. Physicalizations classified according to their use of space and their use of context. Bounded artifacts are confined to a particular position and not meant to be repositioned. Unbounded artifacts are designed for mobility and to move freely across space. Situated and non-situated refer whether the physicalization is displaying information that is contextually relevant in its environment. All images are copyright to their respective owners.

Classifications of physicalizations. A cross in the middle divided this figure into 4 quadrants. The left label for the horizontal axis is labelled as User of Context, the top if Non-situated, and the bottom Situated. The bottom label for the vertical axis is labelled Use of Context. The left is labelled Bounded, and the right Unbounded. The top left quadrant (Non-situated, Bounded) contains 16 square images, each one represents a paper. The top right quadrant (Non-situated, Unbounded) contains 4 square project images.The bottom left quadrant (Situated, Bounded) contains 20 square project images. The bottom right quadrant (Situated, Unbounded) contains 7 square project images.

5.2. Where are physicalizations used or showcased?

Analyzing how physicalizations use space provides insight into how the physicalization leverages its environment. For example, a physicalization’s situatedness—a concept within the visualization (ElSayed et al., 2015) and other relevant HCI communities (e.g., Ubiquitous Computing (Schilit et al., 1994; Weiser, 1991))—describes how information is integrated into its surrounding environment. However, the relationship of the physicalization to its location can also change: physicalizations are either bounded or unbounded to a given location depending on their mobility (Section 3). Bounded artifacts exist confined to a particular place and are not meant to be re-positioned. Unbounded artifacts are designed for mobility and may be relocated by users or natural forces, such as a river flow (e.g., Perovich et al. (Perovich et al., 2021)). Figure 3 shows the distribution of the physicalizations using these two concepts.

Traditional notions of situatedness from visualization focus on how a specific physical location provides context to the presented data to support data comprehension and decision making. While physicalizations are often situated by location (e.g., (von Ompteda, 2019; Perovich et al., 2021)), we found that situatedness also occurs through individuals (i.e., body-embedded physicalizations where the physicalization is integrated with the body) and with respect to activities (i.e., activity-embedded physicalizations where the data reflects a captured or current state of an activity, as discussed in Bressa et al. (Bressa et al., 2022)).

Body-embedded situatedness includes physicalizations that are meant to be worn on the body as either clothing (Howell et al., 2018; Friske et al., 2020), shoes (Nachtigall et al., 2019), accessories (Woźniak et al., 2020), or tattoos (Lo et al., 2016). In many of these cases, the data originates from the wearer through external sensors (e.g., (Howell et al., 2018; Lo et al., 2016)). However, Friske et al. (Friske et al., 2020) encode data through knitting, but the knitter is still the physical referrent for the data. Similarly, Woźniak et al. (Woźniak et al., 2020) used arm and leg bands to provide vibrotactile feedback of the wearer’s elbow bend and body weight imbalance to improve their golfing stance. The arm band is situated with respect to the golfer but the situatedness extends beyond the wearer and into the activity itself, dynamically reacting to the swing in the context of the golfer’s motions. These physicalizations expand the traditional notion that situatedness is siloed to location.

The majority of physicalizations are bound to one location (36/47). Unbounded physicalizations are generally composed of discrete parts that can freely move across space, such as ShapeBots (Suzuki et al., 2019), Zooids (Le Goc et al., 2018), Vital + Morph (Boem and Iwata, 2018), and Chemicals in the Creek (Perovich et al., 2021). For example, Vital + Morph is a shape-changing display allowing families to co-monitor their loved one’s vital signs. Vital + Morph can be disassembled into smaller pieces where each family member can monitor different vital signs at various locations within a city. As an unbounded physicalization, Vital + Morph can be moved freely within and between environments across space. Being distributed across the different family members brings a sense of community and creates a shared understanding, illustrating how physicalizations can be situated beyond their physical environment. These examples reflect and expand on the perspectives of Bressa et al. (Bressa et al., 2022), who state situatedness can occur across space, time, place, activity, and through communities. We additionally found that physicalizations can be situated through the individual (i.e., the body).

Figure 4. Conceptual space that describes the three attributes that contribute to a physicalization’s materiality (i.e., presence in the world (Gross et al., 2014; Wiberg, 2018)).

Conceptual space of the Physicalization Materiality. It has 3 columns. First column on the left is labelled Data Duration. Below it has 3 items, Persistent, Ephemeral, and Permanent. Second column is labelled Embodiment. Below it shows 2 subcategories. First is Technologically-Driven, that includes items Shape-Changing, Embedded Tags, TUI, Swarm Robotics, Light Display, Edibilization, Fluidic Display, and Voxel Prints. Second is Contextually-Driven, that includes Environment, and Activity. The Third column is labelled Material. Below it are Electronic, and Non-electronic.

Figure 5. Eleven examples of physicalization embodiments identified in our corpus. The physical embodiment is categorized as either technologically- or contextually driven. All images are copyright to their respective owners.

Examples of physicalization embodiment. This picture has 2 big boxes. The Left box A Technology-Driven contains 8 categories. Shape-Changing Display has 9 project images. Swarm Robotics has 5 project images. TUI has 4 project images. Fluidic Display has 3 project images. Light has 3 project images. Voxel Prints has 2 project images. Embedded Tag has 1 project image. Edibilization has 1 project image. The Right box B Context Driven has 3 categories. Environment has 9 project images below it. Person has 6 project images. Activity has 4 project images.

5.3. What design decisions contribute to a physicalization’s materiality?

Past work has analyzed physical representation in physicalization through a number of lenses, including pragmatic/artistic (Hogan and Hornecker, 2012; Djavaherpour et al., 2021), epistemological/ontological (Offenhuber, 2020b), and semiotic perspectives (Zhao and Moere, 2008; Offenhuber and Telhan, 2015). To understand physicality within our design space, we investigated each artifacts’ materiality and found three qualities of existence specific to physicalization: embodiment, data duration, and material (Figure 4). We developed a high-level categorization of physicalizations as either technologically-driven or contextually-driven by specifically focusing on how data is manifested physically and temporally. We identified 11 types of embodiment (Figure 5), three classifications of data duration (described in Section 4.2), and distinguish between materials as either electronic or non-electronic.

Embodiment A physicalization’s embodiment provides insight into the designer’s domain expertise, motivations, techniques, and encoding schemes. We categorized embodiment based on how each physicalization manifested data and whether they were technologically-driven or contextually-driven. Technologically-driven physicalizations consisted of platforms where designers leveraged technology to encode data. These platforms included shape-changing displays, swarm robotics, light displays, TUIs, fluidic displays, edibilization, vibrotactile displays, embeddable tags, and voxel prints. Within shape-changing displays, technologies ranged from inflatables to programmable pin arrays to actuated hand-held devices. The materials for technologically-driven physicalizations were chosen based on their performance or use within a particular manufacturing method (e.g., 3D printing, laser cutting, silicone molding).

Physicalizations whose embodiment was contextually-driven represented data through people, environments, or activities. In these cases, physicalizations heavily relied on the context surrounding its use or the user themselves (Lo et al., 2016; Desjardins and Tihanyi, 2019; Howell et al., 2018). These systems emphasized data mapping choices similar to traditional visualization.

Material We categorized the materials used to construct a physicalization as electronic or non-electronic. We intentionally do not use the more common HCI practice of describing an object as either programmable or non-programmable because non-electronic materials can be transformed through craft and other manipulations to become programmable (e.g., the Embroidered Computer (Kurbak, 2018)). We characterize a material as electronic if any manufactured component or collection of manufactured components are used to control the flow of current (microcontrollers, sensors, computers, active and passive components, etc.). Physicalizations that used electronic components can generally be reused with different datasets for new lines of inquiry and analysis (e.g., (Le Goc et al., 2018; Nakagaki et al., 2020; Everitt and Alexander, 2017; Taher et al., 2017)). Physicalizations that had no computational abilities typically targeted a single dataset (e.g., (Thudt et al., 2018; Friske et al., 2020; von Ompteda, 2019; Mor et al., 2020)) or required manual reconstruction of the physicalization by the user.

Data Duration. Physicalizations are often designed for one of three temporal durations: ephemeral, persistent, and permanent (Figure 6). Physicalizations whose data is presented ephemerally collect and render data in real-time and cannot be recorded or recalled. We found that all of these physicalizations (12/47) use sensors to collect and visually stream data. For example, Torrent (Pon et al., 2017) (Figure 6a) augments concert flute music with live water sounds. The water surface turbulence reflects the flutist’ muscle tension throughout the performance. Physicalizations that persist offer users some control over how long data is displayed. These physicalizations often rely on electronics and programming to specify how to retain the data embedded in the physicalization. However, other systems use manual interactions to control data persistence. Venous Materials (Mor et al., 2020) (Figure 6b) uses microfluidics for tangible interactions to represent a dynamic line chart. By directly pressing on a chamber containing the liquid, a user can control the flow of liquid through the microfluidic channel with their finger. People decide how long they would like to see data being displayed before releasing their finger and returning the display to its original state. Physicalizations that display their data permanently are those where the data is permanently embedded or integrated into its physical structure, such as conventional 3D printed objects (Khot et al., 2016; Nissen and Bowers, 2015; Bader et al., 2018; Desjardins and Tihanyi, 2019) or hand-crafted artifacts (Friske et al., 2020; Thudt et al., 2018). Most persistent artifacts did not include any electronic materials. As noted above, such physicalizations cannot decouple materials from data (Friske et al., 2020; Bader et al., 2018), making it difficult to repurpose the artifact for a different dataset without significant effort from the user.

Figure 6. Examples of physicalizations with different data duration (i.e., how long data is displayed through its physical embodiment): (a) Torrent (Pon et al., 2017) displays data ephemerally through its use of water, (b) Venous Materials (Mor et al., 2020) displays data persistently as determined by the user, (c) Data is permanently embedded in Fantibles (Khot et al., 2016).

. Three examples of different data duration in physicalization. Picture A shows 12 flutists with a conductor, and a lighted-up water box next to her on stage in front of an audience. Picture B shows two axes. Vertical line pointing up with hand-written text of Water Usage Projection. The bottom text below the horizontal line shows 1990, 2000, 2010, 2020. The left side of the vertical axis has 3 labels, City A, City B and City C, corresponding to 3 colored coiled fluid channels. An index finger is pressing the coil of City C and the colored fluid line shows a line graph curving up from 1990 toward 2010 and curving down toward 2020. Picture C shows 2 plastic spiky objects with radial patterns. The top green spike is named SPORTOMS that has 20 spikes. The text Start of the game is pointed to the hole in the middle of this spike. Each spike = 5 overs. The small circles on the spikes represent Wickets. The bigger circles at the end of the spikes are Runs. The bottom red object is called FANTOMS. It has a circle with the lengths of spikes representing emotions in Low (no spike), Medium (medium length) and High (longer length).

5.4. What interactions does a physicalization afford?

As discussed in Section 3.2.3, we analyzed how the physicalization’s design features (i) invite people to conduct a particular data interaction (i.e., perceived affordances), (ii) informs what the result of their performed data interaction will be (i.e., feedforward), and (iii) communicate the result of their action (i.e., feedback). Explore (i.e., to interrogate different subsets of the data (Yi et al., 2007)) was the most pervasive interaction technique (12/61). We saw explore applied to everyday objects (e.g., clocks (Lee et al., 2020) cups (Desjardins and Tihanyi, 2019)) and shape-changing displays (Suzuki et al., 2019; Nakagaki et al., 2020). Systems that supported explore allowed people to explore data either directly (akin to direct manipulation in visualization) or indirectly (i.e., akin to dynamic queries (Shneiderman, 1994); Figure 7).

Physicalizations that used direct interactions stimulated the user’s sensory affordances (Hartson, 2003). Most such physicalizations were composed of discrete pieces where each piece is part of the system’s overall interaction scheme. For example, ShapeBots (Suzuki et al., 2019) (Figure 7) is a swarm robotics system where each robot is encoded as a data point within a dataset. With ShapeBots, the researchers intentionally designed the robots to be small, such that the graspable, small pieces invite the user to pick up each piece and interact with the physicalization as a system of parts. ListeningCup (Desjardins and Tihanyi, 2019) also used form to encourage interaction, inviting people to stroke the cup with their hands through its textured surface. In both cases, these systems afford feedforward by primarily appealing to the perceptual-motor skills. ShapeBots’ small graspable components inform people that data can be picked up, and the texture on ListeningCup offers visual cues as to what it might feel like.777For more discussion on sensory-driven design, we refer readers to Hara (Hara, 2007). With the exception of ListeningCups, all physicalizations with direct interactions were made with electronics that responded immediately, changing their state to serve as feedback. For instance, moving a ShapeBot (Suzuki et al., 2019) from one state to another state evoked a change in height, moving a Physikit cube (Houben et al., 2016) from one part of the house to another created change in sensor readings and output in the form of light, vibration, airflow, and movement. ListeningCups, in contrast, provided feedback through its texture which could be experienced through direct touch.

Figure 7. Examples of physicalizations using the explore interaction. Physicalizations using the direct mechanism uses direct manipulation of the object with a focus on sensory affordances. Indirect refers use interface controls (e.g., buttons, sliders, knob) to interact with the data. FluxMarkers (Suzuki et al., 2017) uses speech and touch to explore the data.

Examples of Explore Interaction for physicalization. Direct examples show 3 project pictures: a hand holding a cup, a hand grabbing a small robot on a projected US map, and a hand holding a ball on top of a disk shape. The Indirect examples show 3 project pictures: two hands touching multiple coiled liquid channels: the left hand holding a dial next to the label of population, while the right hand touching one of the different colored balloons controlled by mechanisms with multiple tubes, and a hand touching the corner of the box containing a clock.

Physicalizations that used indirect interaction in contrast presented physical controls that people are familiar with (e.g., buttons, sliders, knobs; Figure 7). People were informed of what the interactive component (e.g., the button or slider) would do through lexical or graphical labels on the physical controls. Similar to direct mechanisms, these physicalizations communicated the result of their action by immediately changing the state of physicalization through use of electronic components.

As physical artifacts, data physicalizations leverage our abilities to manipulate everyday tangible things (e.g., cups, blocks, lamps, clocks) to make sense of intangible data. In addition, data physicalizations with computational components may leverage interface conventions (e.g. sliders, knobs, buttons) to facilitate interaction. Echoing Jansen et al. (Jansen et al., 2015), we observe many parallels between data physicalizations and TUIs. Examining the affordances of physicalizations might enable designers to better leverage TUI principles.

The aforementioned physicalizations employ interaction techniques reflecting known concepts from TUIs, such as physical interactions through graspable elements (Fitzmaurice, 1996) and token+constraint systems (Ullmer et al., 2005). However, interactions within data physicalizations in our corpus—like tangible interactions—cover a wide spectrum of input techniques, from indirect physical controls to more direct coupling of tangible interfaces with digital data (e.g. (Suzuki et al., 2018; Guinness et al., 2019; Veldhuis et al., 2020), see “tangible bits” (Ishii and Ullmer, 1997)) and even pushing towards programmable physical matter (e.g. (Nakagaki et al., 2020; Mor et al., 2020; Le Goc et al., 2016), see “radical atoms” (Ishii et al., 2012)).

With this breadth of different tangible interaction techniques and concepts available, we reiterate that—as interfaces—affordances remain a fundamental consideration while designing physicalizations (Norman, 1988). Affordances for data physicalizations bridge considerations that tie context, interactions, and structure together. For instance, connecting the choice of material and embodiment to interactions can help tailor a physicalization to a particular target audience. FluxMarkers (Suzuki et al., 2017) and Robographics (Guinness et al., 2019) connect electronic swarm robots to the BLV community. These systems both incorporate sound as either input or output and vibrotactile haptic feedback for multi-sensory interaction.

5.5. How do different communities approach physicalization? (And what can we learn from each other?)

Communities involved in physicalization share many of the same values, such as the importance of embodiment and designing specific user tasks. The visualization community has illustrated the importance of understanding a user’s task. A visualization whose focus is to reveal analytical insight for domain experts will be radically different compared to others that depict data for casual use for the general public (Pousman et al., 2007). We contend this logic also extends to physicalizations, and we see an opportunity for different communities to learn from each other by understanding how to best express data with respect to a given task. Physicalizations in TUI and design more deeply explored the value of reflection and agency in design, offering insight into how physicalization might promote self-awareness and prolonged engagement.

5.5.1. Analysis: Importance of Effectiveness

Physicalization is often used for problem-solving and decision-making (i.e., analysis) (Jansen et al., 2015). Analysis is the second most pervasive task (15/47), and other communities can learn from the visualization community specifically in learning how to design systems based on visual effectiveness (e.g., graphical perception) and system flexibility.

As mentioned in Section 2.1, Jansen et al. (Jansen et al., 2015) laid out a research agenda for empirical work that can lead to a better understanding of the effectiveness of physical variables. Ignoring this groundwork may lead to ineffective designs where people are not able to perceive changes in data through the transformations presented by the physicalization. We see this reflected within our corpus where researchers have used volume to communicate changes. Two such examples include Bookly (Ju et al., 2019) and CairnFORM (Daniel et al., 2019) which both encode data using volume. CairnFORM consists of a tower of ten actuatable disks which expanded and retracted in the XY plane. Participants in their user study expressed difficulty in discerning the changes in the diameter of each disk. Bookly is designed with a single motorized disk that expands in the z-direction. Most of their participants did not notice the volumetric change in real-time. Both works discussed how users of their respective systems found it difficult to perceive changes in volume. However, studies in visualization show that people have trouble reasoning over changes in volume and are better at reasoning over other visual variables such as surface area (Jansen and Hornbæk, 2015). Work such as (Sauvé et al., 2021; Jansen and Hornbæk, 2015; López García and Hornecker, 2021) can help researchers design systems that utilize effective physical variables.

Work from the visualization community also highlights the importance of allowing systems to adapt to a given user’s needs. Polysurface (Everitt and Alexander, 2017) provides such flexibility through the actual visual design of their system: the shape-changing display was configurable to three different domain experts’ specific needs. In contrast, Zooids (Le Goc et al., 2018) and EMERGE (Taher et al., 2017) provided flexibility through data interactions. As mentioned in Section 5.1, the depth of the supported data analysis varies depending on the number of interactions available.

5.5.2. Reflection: Importance of Agency

Physicalizations designed for reflective tasks, in contrast, highlight the importance of agency—users should be able to develop personally meaningful data mappings. These customized physicalizations tend to layer meaning-making by encouraging the user to explore and reflect as part of the design process as well as through interacting with the physicalization. A notable example comes from a user within Thudt et al.’s study (Thudt et al., 2018) where tokens were used to visually distil a recipe for an “at a glance” overview. The beads used color to encode ingredients while the number of beads of a particular color encoded how much of each ingredient to use relative to one another. The mapping process inadvertently created a label for the jars of the product generated by each recipe.

Not all physicalizations for reflection provided agency to the users. Data Things (Nissen and Bowers, 2015) and Fantibles (Khot et al., 2016) leveraged users’ personal data such as social media posts but did not include the users in the data mapping process. This exclusion led people to alter their social media feed to reshape their physicalizations and regain their sense of agency during participatory workshops. Even though the design and encoding processes did not intentionally encourage agency, people altered their own behavior to influence the outcome of the physicalization’s form.

6. Discussion

Physicalization draws on traditions and practices from a number of disciplines, including data visualization, TUI, and design. Our survey aligns recent developments across these fields to establish a preliminary design space of 13 key dimensions for physicalization design and development. In developing this design space and synthesizing trends and patterns in physicalization research, we first identify two priorities researchers and designers should consider (Section 6.1 and Section 6.2) and offer opportunities for future work and interdisciplinary collaboration (Section 6.3).

6.1. Considering Context

Context grounds the visual design and functionality behind a data physicalization. We found several instances where the designers of physicalizations were not explicitly clear about the physicalizaton’s context, specifically its target audience, location, and/or task (Section 4.1). While we cannot assert that context was not considered in the design of these artifacts, context is key for both creating effective representations and for connecting relevant bodies of research across disciplines. For example, the tasks people need to perform with their data guide many visualization design workflows and ensure the resulting physicalization communicate attributes of the data that matter to the individual (Sedlmair et al., 2012). Emphasizing context can help researchers identify further end-user considerations that might have not been captured by current research.

Context also offers design guides that may help strengthen physicalizations. For example, physicalizations that aim to support analysis should consider data duration—generally reducing ephemerality and aiming for more persistent or permanent designs to enable people to return to key states that elicit insight into their data (Section 5.3). Designers may also draw on concepts from data provenance to inspire new methods to physicalize data that allow analysts to retrace states of the physicalization over time during analysis (Ragan et al., 2015).

6.2. Defining Data

We used Jansen et al.’s and Hogan and Hornecker’s definitions (Jansen et al., 2015; Hogan and Hornecker, 2012) of physicalization to construct our corpus. However, there is not a clear consensus on what constitutes a physicalization and, more generally, what constitutes data. Offenhuber (Offenhuber, 2020b) notes that there are several definitions of data that vary across bodies of scholarship, including data that stems from the mind (i.e., epistemological). This perspective contrasts the traditional notion of data in the visualization community (i.e., ontological). In works such as Data Agents (Karyda et al., 2021) and Imagining Movie Recommendation Algorithms (Alvarado et al., 2021), mundane objects were used to represent people’s personal data in meaningful ways filtered through the creator’s internal synthesis of the data. These systems do not meet Jansen et al.’s and Hogan and Hornecker’s definitions as data was not physically represented within the physical and visual properties of these data objects or their behaviors. Instead, the artifacts reflect an internal synthesis of data, creating objects whose form is indirectly shaped by data rather than directly representing individual data points.

The reflective focus of many physicalizations and their roots in design and artistic practice challenge traditional definitions of data and data representation from communities like IEEE VIS. One concrete example is the difference in the underlying views of data between the two communities. VIS often takes positivist approach, which views knowledge and data as objectively determined (Meyer and Dykes, 2020). In contrast, the design community frequently employs interpretivist approach where knowledge is viewed as socially constructed and reflected upon. Interpretivist physicalizations consider not only the data’s form, but also how that form may introduce social dynamics that shape interpretation and interaction with the physicalization (e.g., Friske et al. (Friske et al., 2020), Section 5.1). People may see the physicalization as “correct” and objective in ways that may draw into question competing interpretations of the same dataset. However, rather than invalidate a given interpretation and narrative drawn from physicalized data, we call the communities to consider a broadened definition of data and how we approach, consider, and render it.

6.3. Synthesis and Use

Through our design space, we observed multiple dimensions associated with the design of data physicalization. Physicalization designers can use this design space to reason across relevant bodies of research, identify new collaborative opportunities, and explore methods for creating and evaluating physicalizations.

A primary motivation of this paper is to bridge knowledge from visualization, TUI, and design in order to provide a cohesive lens of each relevant intellectual community’s set of approaches, theories, and terminologies. The presented design space is intended to help designers and researchers be more cognizant of knowledge and practices emerging in other communities (Section 5.5). This need is especially pertinent as our conducted survey shows data physicalization researchers come from various backgrounds (as seen in Section 4), yet these papers seldom draw substantively on the literature of the other communities. While the lack of cross-referencing does not itself imply a lack of interchange, we also found a lack of cross-pollination of ideas, techniques, and methods. Our survey aims to bridge by presenting a shared context for exploring vital facets of physicalization design, including the value of physicalizations as means for for data communication in visualization, for expressivity and reflection in design, and for understanding and demonstrating novel techniques and material use in TUI.

The interdisciplinary nature of data physicalization also opens new opportunities for closer collaborations among relevant communities. As shown in Section 5.5, we argue for closer collaboration particularly from the standpoint of how cross-pollination may aid in generating new data physicalization artifacts and evaluating existing ones. This interdisciplinary cross-over can happen at both a low-level (e.g., integrating data visualization knowledge regarding visual perception) and at a higher-level (e.g., considering how data is integrated into a physicalization’s overall form). Future work generating and evaluating data physicalization may benefit from a research through design approach, borrowing on related methods employed in each of the surveyed fields.

6.3.1. Research through Design

Research through design is prevalent in visualization (Sedlmair et al., 2012; Meyer and Dykes, 2020) (often instantiated through design studies) and the broader HCI community (Zimmerman et al., 2007) because of its emphasis on synthesizing knowledge into meaningful artifacts for both research and practice. Each data physicalization artifact is (un)consciously a product of different community perspectives and values. Collaborative applications of research through design in physicalization offer an opportunity to investigate new techniques through the lens provided by each community. However, successful applications of these methods and integration of community perspectives requires considering how we can be more intentional with the physicalizations we produce to enable the blending of ideas and values. We encourage designers and researchers to refer to the criteria laid out by Zimmerman et al. (Zimmerman et al., 2007) and Meyer and Dykes (Meyer and Dykes, 2020) for integrating rigorous design into their research practice. These frameworks emphasize synthesizing diverse considerations for constructing and evaluating research artifacts in the real world. From our analysis of the data physicalization corpus covered in this paper, two criteria stand out to us in particular:

Invention. One example of generating new design possibilities is through the lens of invention (Zimmerman et al., 2007). For example, a multidisiplinary lens on interactive physicalization may offer innovations for more sustainable interaction design. Section 5.3 and Section 4.3 show that most electronics physicalizations provide easy updates to data representations due to the automation and computation capabilities offered by electronic devices. These properties make the physicalization reusable with multiple datasets and often increase their data and task scalability. However, electronic physicalizations are often expensive, complex, and ecologically unsustainable (Everitt and Alexander, 2017). Our survey shows that electronic materials are not the only technology that can provide automation and computation. Mechanical, chemical, and biological actuated physicalization offer many of the same computational benefits of electronic physicalization while potentially being more ecologically friendly. For example, Venous Materials (Mor et al., 2020) provide interaction based on microfludiics, and Printed Paper Markers (Zheng et al., 2020) employ kirigami techniques for actuation design where paper acts as buttons with tactile feedback. Future work in data physicalization can consider new or emerging domains relevant to HCI, such as the physical and biological sciences, to investigate new interaction paradigms and design dimensions.

Extensibility/Resonance. Both Zimmerman et al. (Zimmerman et al., 2007) and Meyer and Dykes (Meyer and Dykes, 2020) contend that the artifacts produced through design studies should communicate implicit, hidden knowledge and build upon existing knowledge. Applying this perspective across multiple communities may help designers better comprehend and leverage embodiment in physicalization. As discussed in Section 5.3, we identified 11 types of physicalization embodiment. This variety indicates a growing exploration of how to materialize data. However, there is a lack of knowledge transfer among each embodiment type (i.e., technology-driven, contextually-driven) and category (e.g., shape-changing display, swarm robotics; Figure 5). Throughout our analysis, we found that researchers seldom explicitly situate their artifact within the bigger landscape of data physicalization. This approach can silo physicalizations from each other, limiting our abilities to connect between different innovations and subdisciplines. We anticipate that our design space can help remedy this siloing by providing a shared physical representation vocabulary. Moving forward, we encourage researchers and designers to use this design space as a starting point to draw inspiration for new physicalizations and transfer knowledge produced from their artifact to general data physicalization knowledge.

Collectively, our work suggests the need for the data visualization, TUI, and design communities to engage in a collaborative dialogue around design motivations, methods, and techniques for physicalization development. Such cross-disciplinary collaboration will lead to mutually beneficial innovation that will help advance the field of data physicalization. We hope this paper serves as a call to action for members of these communities to recognize and take advantage of potential synergies among the relevant bodies of knowledge to drive novel approaches to physicalization.

6.4. Limitations

Grounded theory requires data collection in an environment constructed by the researchers. Although measures were put in place to maximize credibility and repeatability, code application depends on the definitions and perspectives of the coders. Different coders may result in different findings. Papers within the corpus were not always explicit with respect to each dimension and were sometimes non-exhaustive (e.g., locations or audiences were not always described), introducing additional uncertainty. To mitigate potential bias, the authors consist of researchers from various fields, and we used peer debriefing to ensure that conclusions were grounded in the data. Physicalizations are diverse not only in their structure but use and interpretation of data (Section 6.2). This diversity created challenges in determining whether a paper belonged within the corpus. While we aim for coverage of a range of HCI disciplines, our survey is not intended to be comprehensive. Physicalizations exist beyond research articles (e.g., (Segal, ; Weinberg, )), which we have not covered within our corpus. This choice limits the scope of our survey to artifacts with an emphasis on human interactions with data as a product of technological innovation (e.g., materials engineering, electronic interactions, data analytics, craft). Communities of practice outside of HCI, such as art, may offer alternative frameworks for reasoning about physicalization design and complement the dimensions of our design space. Future work should investigate and refine the design space through the integration of a broader range of knowledge.

7. Conclusion

We surveyed 47 physicalizations to understand common design elements of physicalization research across visualization and HCI. We used these elements to characterize thirteen dimensions of physicalization work. Our survey revealed that physicalizations offer solutions for a diverse range of tasks with objectives and designs that extend beyond the analytical, expert-driven focus of systems from conventional visualization research, such as emphasizing reflection and public engagement. These innovations arise from the deeply cross-disciplinary traditions of the field. We reflect on these traditions and gaps in the design space to identify key open challenges in physicalization research that require rich, collaborative perspectives to address. We hope that this perspective can proactively bridge the diverse knowledge and techniques that physicalization research inspires.

This research is sponsored in part by the U.S. National Science Foundation through grants IIS-2040489 & STEM+C 1933915.


  • O. Alvarado, V. Vanden Abeele, D. Geerts, F. Gutiérrez, and K. Verbert (2021) Exploring tangible algorithmic imaginaries in movie recommendations. In Proceedings of the Fifteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’21, New York, NY, USA. External Links: ISBN 9781450382137, Link, Document Cited by: §6.2.
  • I. Aslan, A. Seiderer, C. T. Dang, S. Rädler, and E. André (2020) PiHearts: resonating experiences of self and others enabled by a tangible somaesthetic design. In Proceedings of the 2020 International Conference on Multimodal Interaction, ICMI ’20, New York, NY, USA, pp. 433–441. External Links: Document, ISBN 9781450375818, Link Cited by: §2.1, §4.2.
  • C. Bader, D. Kolb, J. C. Weaver, S. Sharma, A. Hosny, J. Costa, and N. Oxman (2018) Making data matter: voxel printing for the digital fabrication of data across scales and domains. Science advances 4 (5), pp. eaas8652. Cited by: §3.1, §5.3.
  • A. Batch, B. Patnaik, M. Akazue, and N. Elmqvist (2020) Scents and sensibility: evaluating information olfactation. pp. 1–14. External Links: ISBN 9781450367080, Link Cited by: item 4.
  • M. Beaudouin-Lafon (2004) Designing interaction, not interfaces. In Proceedings of the Working Conference on Advanced Visual Interfaces, AVI ’04, New York, NY, USA, pp. 15–22. External Links: ISBN 1581138679, Link, Document Cited by: §1.
  • A. Boem and H. Iwata (2018) “It’s like holding a human heart”: the design of vital+ morph, a shape-changing interface for remote monitoring. AI & SOCIETY 33 (4), pp. 599–619. Cited by: §2.1, §3.1, §4.1, §4.1, §5.1, §5.2.
  • S. Braley, C. Rubens, T. Merritt, and R. Vertegaal (2018) GridDrones: a self-levitating physical voxel lattice for interactive 3d surface deformations. In Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology, UIST ’18, New York, NY, USA, pp. 87–98. External Links: Document, ISBN 9781450359481, Link Cited by: item 1.
  • M. Brehmer and T. Munzner (2013) A multi-level typology of abstract visualization tasks. IEEE transactions on visualization and computer graphics 19 (12), pp. 2376–2385. Cited by: §2.2.
  • N. Bressa, H. Korsgaard, A. Tabard, S. Houben, and J. Vermeulen (2022) What’s the situation with situated visualization? a survey and perspectives on situatedness. IEEE Transactions on Visualization & Computer Graphics 28 (01), pp. 107–117. External Links: ISSN 1941-0506, Document Cited by: §2.2, §5.2, §5.2.
  • Card,S., Mackinlay,J.D., and Shneiderman,B. (1999) Readings in information visualization: using vision to think. Technical Communication Quarterly 9 (3), pp. 347–351. Cited by: §2.1, §4.1.
  • M. Chalmers (2004) A historical view of context. Computer Supported Cooperative Work (CSCW) 13 (3-4), pp. 223–247. Cited by: §3.2.1.
  • S. Claes and A. V. Moere (2015) The role of tangible interaction in exploring information on public visualization displays. In Proceedings of the 4th International Symposium on Pervasive Displays, PerDis ’15, New York, NY, USA, pp. 201–207. External Links: ISBN 9781450336086, Link, Document Cited by: §4.1.
  • M. Daniel, G. Rivière, and N. Couture (2019) CairnFORM: a shape-changing ring chart notifying renewable energy availability in peripheral locations. In Proceedings of the Thirteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’19, New York, NY, USA, pp. 275–286. External Links: Document, ISBN 9781450361965, Link Cited by: §4.1, §4.1, §5.5.1.
  • [14] P. Dragicevic and Y. Jansen Ammassalik wooden maps. Note: Accessed: 2021-08-29 Cited by: §1.
  • [15] P. Dragicevic and Y. Jansen Marshall islands stick charts. Note: Accessed: 2021-08-29 Cited by: §1.
  • A. Desjardins and T. Tihanyi (2019) ListeningCups: a case of data tactility and data stories. In Proceedings of the 2019 on Designing Interactive Systems Conference, DIS ’19, New York, NY, USA, pp. 147–160. External Links: ISBN 9781450358507, Link, Document Cited by: §5.1, §5.3, §5.3, §5.4, §5.4.
  • H. Djavaherpour, F. Samavati, A. Mahdavi-Amiri, F. Yazdanbakhsh, S. Huron, R. Levy, Y. Jansen, and L. Oehlberg (2021) Data to physicalization: a survey of the physical rendering process. Computer Graphics Forum 40 (3), pp. 569–598. External Links: Document, Link Cited by: §1, §3.2.2, §5.3.
  • P. Dourish (2001) Where the action is: the foundations of embodied interaction. MIT Press, Cambridge, MA, USA. External Links: ISBN 0262041960 Cited by: footnote 4.
  • P. Dragicevic, Y. Jansen, and A. Vande Moere (2020) Data physicalization. In Handbook of Human Computer Interaction, pp. 1–51. External Links: ISBN 978-3-319-27648-9, Document, Link Cited by: §1, §2.2.
  • N. ElSayed, B. Thomas, K. Marriott, J. Piantadosi, and R. Smith (2015) Situated analytics. In 2015 Big Data Visual Analytics (BDVA), pp. 1–8. Cited by: §5.2.
  • C. Elsden, D. Kirk, M. Selby, and C. Speed (2015) Beyond personal informatics: designing for experiences with data. In Proceedings of the 33rd Annual ACM Conference Extended Abstracts on Human Factors in Computing Systems, CHI EA ’15, New York, NY, USA, pp. 2341–2344. External Links: ISBN 9781450331463, Link, Document Cited by: §5.1.
  • O. Ettehadi, F. Anderson, A. Tindale, and S. Somanath (2021) Documented: embedding information onto and retrieving information from 3d printed objects. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, New York, NY, USA. External Links: ISBN 9781450380966, Link Cited by: §4.1, §4.2.
  • A. Everitt and J. Alexander (2017) PolySurface: a design approach for rapid prototyping of shape-changing displays using semi-solid surfaces. In Proceedings of the 2017 Conference on Designing Interactive Systems, DIS ’17, New York, NY, USA, pp. 1283–1294. External Links: Document, ISBN 9781450349222, Link Cited by: §2.1, §5.3, §5.5.1, §6.3.1.
  • G. W. Fitzmaurice (1996) Graspable user interfaces. Ph.D. Thesis, University of Toronto, CAN. Note: AAINN18871 External Links: ISBN 061218871X Cited by: §5.4.
  • M. Friske, J. Wirfs-Brock, and L. Devendorf (2020) Entangling the roles of maker and interpreter in interpersonal data narratives: explorations in yarn and sound. In Proceedings of the 2020 ACM Designing Interactive Systems Conference, New York, NY, USA, pp. 297–310. External Links: ISBN 9781450369749, Link Cited by: §2.1, §2.1, §5.1, §5.2, §5.3, §5.3, §6.2.
  • A. Gendreau Chakarov, Q. Biddy, J. Jacobs, M. Recker, and T. Sumner (2020) Opening the black box: investigating student understanding of data displays using programmable sensor technology. In Proceedings of the 2020 ACM Conference on International Computing Education Research, ICER ’20, New York, NY, USA, pp. 291–301. External Links: Document, ISBN 9781450370929, Link Cited by: §4.1, §4.2.
  • P. Gourlet and T. Dassé (2017) <I>cairn</i>: a tangible apparatus for situated data collection, visualization and analysis. In Proceedings of the 2017 Conference on Designing Interactive Systems, DIS ’17, New York, NY, USA, pp. 247–258. External Links: Document, ISBN 9781450349222, Link Cited by: §4.1.
  • S. Gross, J. Bardzell, and S. Bardzell (2014) Structures, forms, and stuff: the materiality and medium of interaction. Personal and Ubiquitous Computing 18 (3), pp. 637–649. Cited by: §2.1, §3.2.2, Figure 4.
  • D. Guinness, A. Muehlbradt, D. Szafir, and S. K. Kane (2019) RoboGraphics: dynamic tactile graphics powered by mobile robots. In The 21st International ACM SIGACCESS Conference on Computers and Accessibility, ASSETS ’19, New York, NY, USA, pp. 318–328. External Links: Document, ISBN 9781450366762, Link Cited by: §5.1, §5.4, §5.4.
  • K. Hara (2007) Designing design. Lars Müller Publishers, . Cited by: footnote 7.
  • R. Hartson (2003) Cognitive, physical, sensory, and functional affordances in interaction design. Behaviour & Information Technology 22 (5), pp. 315–338. External Links: Document, Link Cited by: §5.4.
  • T. Hogan and E. Hornecker (2012) How does representation modality affect user-experience of data artifacts?. In Haptic and Audio Interaction Design, C. Magnusson, D. Szymczak, and S. Brewster (Eds.), Berlin, Heidelberg, pp. 141–151. External Links: ISBN 978-3-642-32796-4 Cited by: §2.1, §2.2, §5.3, §6.2.
  • E. Hornecker (2005) A design theme for tangible interaction: embodied facilitation. In ECSCW 2005, H. Gellersen, K. Schmidt, M. Beaudouin-Lafon, and W. Mackay (Eds.), Dordrecht, pp. 23–43. External Links: ISBN 978-1-4020-4023-8 Cited by: §3.2.2.
  • S. Houben, C. Golsteijn, S. Gallacher, R. Johnson, S. Bakker, N. Marquardt, L. Capra, and Y. Rogers (2016) Physikit: data engagement through physical ambient visualizations in the home. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems, pp. 1608–1619. External Links: ISBN 9781450333627, Link Cited by: §4.1, §5.1, §5.4.
  • N. Howell, L. Devendorf, T. A. Vega Gálvez, R. Tian, and K. Ryokai (2018) Tensions of data-driven reflection: a case study of real-time emotional biosensing. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, pp. 1–13. External Links: ISBN 9781450356206, Link Cited by: §2.1, §4.1, §4.1, §4.1, §5.1, §5.2, §5.3.
  • C. Hull and W. Willett (2017) Building with data: architectural models as inspiration for data physicalization. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, New York, NY, USA, pp. 1217–1264. External Links: ISBN 9781450346559, Link Cited by: §2.2.
  • H. Ishii, D. Lakatos, L. Bonanni, and J. Labrune (2012) Radical atoms: beyond tangible bits, toward transformable materials. Interactions 19 (1), pp. 38–51. External Links: ISSN 1072-5520, Link, Document Cited by: §5.4.
  • H. Ishii and B. Ullmer (1997) Tangible bits: towards seamless interfaces between people, bits and atoms. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems, CHI ’97, New York, NY, USA, pp. 234–241. External Links: Document, ISBN 0897918029, Link Cited by: §3.2.2, §5.4.
  • Y. Jansen, P. Dragicevic, and J. Fekete (2013) Evaluating the efficiency of physical visualizations. pp. 2593–2602. External Links: ISBN 9781450318990, Link Cited by: §3.1.
  • Y. Jansen, P. Dragicevic, P. Isenberg, J. Alexander, A. Karnik, J. Kildal, S. Subramanian, and K. Hornbæk (2015) Opportunities and challenges for data physicalization. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, pp. 3227–3236. External Links: ISBN 9781450331456, Link Cited by: §1, §1, §2.1, §2.1, item 2, §5.1, §5.4, §5.5.1, §5.5.1, §6.2.
  • Y. Jansen and K. Hornbæk (2015) A psychophysical investigation of size as a physical variable. IEEE transactions on visualization and computer graphics 22 (1), pp. 479–488. Cited by: §2.1, §5.5.1.
  • Y. Jansen and K. Hornbæk (2016) A psychophysical investigation of size as a physical variable. IEEE Transactions on Visualization and Computer Graphics 22 (1), pp. 479–488. External Links: Document Cited by: item 3.
  • A. Jofre, S. Szigeti, S. T. Keller, L. Dong, D. Czarnowski, F. Tomé, and S. Diamond (2015) A tangible user interface for interactive data visualization. In Proceedings of the 25th Annual International Conference on Computer Science and Software Engineering, CASCON ’15, USA, pp. 244–247. Cited by: item 2.
  • Joosten,Jan (1996) Denise schmandt-besserat, how writing came about. University of Texas Press, Austin, TX, USA. External Links: ISBN 0035-2403 Cited by: §1.
  • S. Ju, K. Lee, S. Kim, and Y. Park (2019) Bookly: an interactive everyday artifact showing the time of physically accumulated reading activity. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, pp. 1–8. External Links: ISBN 9781450359702, Link Cited by: §4.1, §4.1, §4.1, §4.1, §5.1, §5.5.1.
  • M. Kaninsky, S. Gallacher, and Y. Rogers (2018) Confronting people’s fears about bats: combining multi-modal and environmentally sensed data to promote curiosity and discovery. In Proceedings of the 2018 Designing Interactive Systems Conference, DIS ’18, New York, NY, USA, pp. 931–943. External Links: Document, ISBN 9781450351980, Link Cited by: item 2.
  • H. G. Kaper, E. Wiebel, and S. Tipei (1999) Data sonification and sound visualization. Computing in science & engineering 1 (4), pp. 48–58. Cited by: item 4.
  • M. Karyda, E. D. Mekler, and A. Lucero (2021) Data agents: promoting reflection through meaningful representations of personal data in everyday life. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, New York, NY, USA. External Links: ISBN 9781450380966, Link Cited by: §6.2.
  • M. Karyda, D. Wilde, and M. G. Kjærsgaard (2020) Narrative physicalization: supporting interactive engagement with personal data. IEEE Computer Graphics and Applications 41 (1), pp. 74–86. Cited by: item 5.
  • R. A. Khot, D. Aggarwal, R. Pennings, L. Hjorth, and F. ’. Mueller (2017) <I>edipulse</i>: investigating a playful approach to self-monitoring through 3d printed chocolate treats. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, pp. 6593–6607. External Links: ISBN 9781450346559, Link Cited by: §4.1.
  • R. A. Khot, J. Andres, J. Lai, J. von Kaenel, and F. ’. Mueller (2016) <I>fantibles</i>: capturing cricket fan’s story in 3d. In Proceedings of the 2016 ACM Conference on Designing Interactive Systems, DIS ’16, New York, NY, USA, pp. 883–894. External Links: Document, ISBN 9781450340311, Link Cited by: §4.1, §4.1, §4.1, §4.1, Figure 6, §5.3, §5.5.2.
  • K. Krippendorff (2004) Content analysis: an introduction to its methodology. Content Analysis: An Introduction to Its Methodology, Sage. External Links: ISBN 9780761915454, LCCN 2003014200, Link Cited by: §3.3.
  • S. Kuchera (2018) The weavers and their information webs: steganography in the textile arts. Ada: A Journal of Gender, New Media, and Technology 13. External Links: Link Cited by: §1.
  • E. Kurbak (2018) Stitching worlds: exploring textiles and electronics. Revolver Publishing, Berlin. Cited by: §1, §5.3.
  • M. Le Goc, L. H. Kim, A. Parsaei, J. Fekete, P. Dragicevic, and S. Follmer (2016) Zooids: building blocks for swarm user interfaces. In Proceedings of the 29th Annual Symposium on User Interface Software and Technology, UIST ’16, New York, NY, USA, pp. 97–109. External Links: Document, ISBN 9781450341899, Link Cited by: §5.4, footnote 3.
  • M. Le Goc, C. Perin, S. Follmer, J. Fekete, and P. Dragicevic (2018) Dynamic composite data physicalization using wheeled micro-robots. IEEE Transactions on Visualization and Computer Graphics 25 (1), pp. 737–747. Cited by: §4.1, §5.1, §5.2, §5.3, §5.5.1, footnote 3.
  • K. Lee, S. Ju, T. Dzhoroev, G. Goh, M. Lee, and Y. Park (2020) DayClo: an everyday table clock providing interaction with personal schedule data for self-reflection. In Proceedings of the 2020 ACM Designing Interactive Systems Conference, pp. 1793–1806. External Links: ISBN 9781450369749, Link Cited by: §4.1, §4.1, §5.4.
  • K. Lee, B. Kim, J. Kim, H. Hong, and Y. Park (2021) ADIO: an interactive artifact physically representing the intangible digital audiobook listening experience in everyday living spaces. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, New York, NY, USA. External Links: ISBN 9781450380966, Link Cited by: §4.1.
  • J. Lo, D. J. L. Lee, N. Wong, D. Bui, and E. Paulos (2016) Skintillates: designing and creating epidermal interactions. In Proceedings of the 2016 ACM Conference on Designing Interactive Systems, DIS ’16, New York, NY, USA, pp. 853–864. External Links: Document, ISBN 9781450340311, Link Cited by: §4.1, §4.2, §5.2, §5.3.
  • D. Lockton, T. Zea-Wolfson, J. Chou, Y. (. Song, E. Ryan, and C. Walsh (2020) Sleep ecologies: tools for snoozy autoethnography. pp. 1579–1591. External Links: ISBN 9781450369749, Link Cited by: item 5.
  • I. López García and E. Hornecker (2021) Scaling data physicalization – how does size influence experience?. In Proceedings of the Fifteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’21, New York, NY, USA. External Links: Document, ISBN 9781450382137, Link Cited by: item 3, §5.5.1.
  • T. M. Madhyastha and D. A. Reed (1995) Data sonification: do you see what i hear?. IEEE Software 12 (2), pp. 45–56. Cited by: item 4.
  • R. Manshaei, U. Mayat, A. Tarun, S. DeLong, D. Chiang, J. Digregorio, S. Khayyer, A. Gupta, M. Kyan, and A. Mazalek (2019)

    Tangible tensors: an interactive system for grasping trends in biological systems modeling

    In Proceedings of the 2019 on Creativity and Cognition, C&C ’19, New York, NY, USA, pp. 41–52. External Links: Document, ISBN 9781450359177, Link Cited by: item 2.
  • S. Marco and A. Gutierrez-Galvez (2012) Signal and data processing for machine olfaction and chemical sensing: a review. IEEE Sensors Journal 12 (11), pp. 3189–3214. Cited by: item 4.
  • D. Menheere, E. van Hartingsveldt, M. Birkebæk, S. Vos, and C. Lallemand (2021) Laina: dynamic data physicalization for slow exercising feedback. In Designing Interactive Systems Conference 2021, DIS ’21, New York, NY, USA, pp. 1015–1030. External Links: ISBN 9781450384766, Link, Document Cited by: §4.1, §4.1.
  • M. Meyer and J. Dykes (2020) Criteria for rigor in visualization design study. IEEE Transactions on Visualization and Computer Graphics 26 (1), pp. 87–97. External Links: Document Cited by: §6.2, §6.3.1, §6.3.1.
  • H. Mor, T. Yu, K. Nakagaki, B. H. Miller, Y. Jia, and H. Ishii (2020) Venous materials: towards interactive fluidic mechanisms. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, pp. 1–14. External Links: ISBN 9781450367080, Link Cited by: §2.1, §4.2, Figure 6, §5.3, §5.3, §5.4, §6.3.1.
  • Munzner,Tamara (2015) Visualization analysis & design. CRC Press Taylor & Francis Group, Boca Raton, FL. Cited by: §2.1, §4.1.
  • T. Nachtigall, O. Tomico, R. Wakkary, and P. van Dongen (2019) Encoding materials and data for iterative personalization. pp. 1–12. External Links: ISBN 9781450359702, Link Cited by: §3.1, §4.1, §5.2.
  • K. Nakagaki, S. Follmer, and H. Ishii (2015) LineFORM: actuated curve interfaces for display, interaction, and constraint. In Proceedings of the 28th Annual Symposium on User Interface Software and Technology, UIST ’15, New York, NY, USA, pp. 333–339. External Links: Document, ISBN 9781450337793, Link Cited by: §4.1.
  • K. Nakagaki, Y. (. Liu, C. Nelson-Arzuaga, and H. Ishii (2020) TRANS-dock: expanding the interactivity of pin-based shape displays by docking mechanical transducers. In Proceedings of the Fourteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’20, New York, NY, USA, pp. 131–142. External Links: Document, ISBN 9781450361071, Link Cited by: §2.1, §4.1, §5.3, §5.4, §5.4.
  • S. Nath (2020) Hear her fear: data sonification for sensitizing society on crime against women in india. In IndiaHCI ’20: Proceedings of the 11th Indian Conference on Human-Computer Interaction, IndiaHCI 2020, New York, NY, USA, pp. 86–91. External Links: Document, ISBN 9781450389440, Link Cited by: item 4.
  • B. Nissen and J. Bowers (2015) Data-things: digital fabrication situated within participatory data translation activities. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, pp. 2467–2476. External Links: ISBN 9781450331456, Link Cited by: §4.1, §4.1, §5.1, §5.3, §5.5.2.
  • D. A. Norman (1988) The psychology of everyday things. Basic Books. Cited by: §5.4.
  • D. Offenhuber and O. Telhan (2015) Indexical visualization–the data-less information display. In Ubiquitous Computing, Complexity, and Culture, pp. 288–302. Cited by: §1, §2.2, §5.3.
  • D. Offenhuber (2020a) Data by proxy — material traces as autographic visualizations. IEEE Transactions on Visualization and Computer Graphics 26 (1), pp. 98–108. External Links: Document Cited by: §2.2.
  • D. Offenhuber (2020b) What we talk about when we talk about data physicality. IEEE Computer Graphics and Applications 40 (6), pp. 25–37. External Links: Document Cited by: §1, §2.1, §2.1, §5.3, §6.2.
  • M. J. Page, J. E. McKenzie, P. M. Bossuyt, I. Boutron, T. C. Hoffmann, C. D. Mulrow, L. Shamseer, J. M. Tetzlaff, E. A. Akl, S. E. Brennan, R. Chou, J. Glanville, J. M. Grimshaw, A. Hróbjartsson, M. M. Lalu, T. Li, E. W. Loder, E. Mayo-Wilson, S. McDonald, L. A. McGuinness, L. A. Stewart, J. Thomas, A. C. Tricco, V. A. Welch, P. Whiting, and D. Moher (2021) The prisma 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372. External Links: Document, Link, Cited by: Figure 2.
  • G. Pahl and W. Beitz (1996) Embodiment design. pp. 199–403. Cited by: footnote 4.
  • J. Pepping, S. Scholte, M. van Wijland, M. de Meij, G. Wallner, and R. Bernhaupt (2020) Motiis: fostering parents’ awareness of their adolescents emotional experiences during gaming. In Proceedings of the 11th Nordic Conference on Human-Computer Interaction: Shaping Experiences, Shaping Society, NordiCHI ’20, New York, NY, USA. External Links: Document, ISBN 9781450375795, Link Cited by: §2.1, §4.1, §4.1, §4.2.
  • L. J. Perovich, S. A. Wylie, and R. Bongiovanni (2021) Chemicals in the creek: designing a situated data physicalization of open government data with the community. IEEE Transactions on Visualization and Computer Graphics 27 (2), pp. 913–923. External Links: Document Cited by: §4.1, §4.1, §4.2, §5.2, §5.2, §5.2.
  • A. Pon, E. Pattison, L. Fyfe, L. Radford, and S. Carpendale (2017) <I>torrent</i>: integrating embodiment, physicalization and musification in music-making. In Proceedings of the Eleventh International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’17, New York, NY, USA, pp. 209–216. External Links: Document, ISBN 9781450346764, Link Cited by: §2.1, §4.1, §4.1, §4.1, §4.1, §4.2, Figure 6, §5.3.
  • Z. Pousman, J. Stasko, and M. Mateas (2007) Casual information visualization: depictions of data in everyday life. IEEE Transactions on Visualization and Computer Graphics 13 (6), pp. 1145–1152. External Links: Document Cited by: §5.5.
  • F. Quintal, C. Jorge, V. Nisi, and N. Nunes (2016) Watt-i-see: a tangible visualization of energy. In Proceedings of the International Working Conference on Advanced Visual Interfaces, AVI ’16, New York, NY, USA, pp. 120–127. External Links: Document, ISBN 9781450341318, Link Cited by: §2.1, §4.1, §4.2.
  • E. D. Ragan, A. Endert, J. Sanyal, and J. Chen (2015) Characterizing provenance in visualization and data analysis: an organizational framework of provenance types and purposes. IEEE transactions on visualization and computer graphics 22 (1), pp. 31–40. Cited by: §6.1.
  • R. G. Raidou, M. E. Gröller, and H. Wu (2020) Slice and dice: a physicalization workflow for anatomical edutainment. Computer Graphics Forum 39 (7), pp. 623–634. External Links: Document, Link, Cited by: §3.1.
  • E. Sabinson, I. Pradhan, and K. Evan Green (2021) Plant-human embodied biofeedback (pheb): a soft robotic surface for emotion regulation in confined physical space. In Proceedings of the Fifteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’21, New York, NY, USA. External Links: ISBN 9781450382137, Link, Document Cited by: §4.1, §4.2.
  • D. R. Sahoo, T. Nakamura, A. Marzo, T. Omirou, M. Asakawa, and S. Subramanian (2016) JOLED: a mid-air display based on electrostatic rotation of levitated janus objects. In Proceedings of the 29th Annual Symposium on User Interface Software and Technology, UIST ’16, New York, NY, USA, pp. 437–448. External Links: Document, ISBN 9781450341899, Link Cited by: item 1.
  • A. Sarikaya, M. Gleicher, and D. A. Szafir (2018) Design factors for summary visualization in visual analytics. Computer Graphics Forum 37 (3), pp. 145–156. External Links: Document, Link, Cited by: §4.1.
  • K. Sauvé, S. Bakker, and S. Houben (2020a) Econundrum: visualizing the climate impact of dietary choice through a shared data sculpture. In Proceedings of the 2020 ACM Designing Interactive Systems Conference, pp. 1287–1300. External Links: ISBN 9781450369749, Link Cited by: §4.1.
  • K. Sauvé, S. Bakker, N. Marquardt, and S. Houben (2020b) LOOP: exploring physicalization of activity tracking data. In Proceedings of the 11th Nordic Conference on Human-Computer Interaction: Shaping Experiences, Shaping Society, NordiCHI ’20, New York, NY, USA. External Links: Document, ISBN 9781450375795, Link Cited by: §4.1, §4.1, §4.1.
  • K. Sauvé, D. Verweij, J. Alexander, and S. Houben (2021) Reconfiguration strategies with composite data physicalizations. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, CHI ’21, New York, NY, USA. External Links: Document, ISBN 9781450380966, Link Cited by: §5.5.1.
  • B. Schilit, N. Adams, and R. Want (1994) Context-aware computing applications. In 1994 First Workshop on Mobile Computing Systems and Applications, pp. 85–90. Cited by: §5.2.
  • M. Sedlmair, M. Meyer, and T. Munzner (2012) Design study methodology: reflections from the trenches and the stacks. IEEE transactions on visualization and computer graphics 18 (12), pp. 2431–2440. Cited by: §2.2, §6.1, §6.3.1.
  • [95] A. Segal Note: 2021-09-09 Cited by: §6.4.
  • B. Shneiderman (1994) Dynamic queries for visual information seeking. IEEE software 11 (6), pp. 70–77. Cited by: §5.4.
  • S. Stusak, M. Hobe, and A. Butz (2016) If your mind can grasp it, your hands will help. In Proceedings of the TEI ’16: Tenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’16, New York, NY, USA, pp. 92–99. External Links: Document, ISBN 9781450335829, Link Cited by: item 3.
  • R. Suzuki, J. Kato, M. D. Gross, and T. Yeh (2018) Reactile: programming swarm user interfaces through direct physical manipulation. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, pp. 1–13. External Links: ISBN 9781450356206, Link Cited by: §4.1, §5.4.
  • R. Suzuki, A. Stangl, M. D. Gross, and T. Yeh (2017) FluxMarker: enhancing tactile graphics with dynamic tactile markers. In Proceedings of the 19th International ACM SIGACCESS Conference on Computers and Accessibility, ASSETS ’17, New York, NY, USA, pp. 190–199. External Links: Document, ISBN 9781450349260, Link Cited by: Figure 7, §5.1, §5.4.
  • R. Suzuki, C. Zheng, Y. Kakehi, T. Yeh, E. Y. Do, M. D. Gross, and D. Leithinger (2019) ShapeBots: shape-changing swarm robots. In Proceedings of the 32nd Annual ACM Symposium on User Interface Software and Technology, UIST ’19, New York, NY, USA, pp. 493–505. External Links: Document, ISBN 9781450368162, Link Cited by: §4.1, §5.2, §5.4, §5.4.
  • F. Taher, Y. Jansen, J. Woodruff, J. Hardy, K. Hornbæk, and J. Alexander (2017) Investigating the use of a dynamic physical bar chart for data exploration and presentation. IEEE Transactions on Visualization and Computer Graphics 23 (1), pp. 451–460. External Links: Document Cited by: §2.1, §4.1, §5.1, §5.3, §5.5.1.
  • C. E. Tejada, R. Ramakers, S. Boring, and D. Ashbrook (2020) AirTouch: 3d-printed touch-sensitive objects using pneumatic sensing. In Proceedings of the 2020 CHI Conference on Human Factors in Computing SystemsProceedings of the 2020 ACM Designing Interactive Systems ConferenceProceedings of the SIGCHI Conference on Human Factors in Computing SystemsProceedings of the 2020 CHI Conference on Human Factors in Computing SystemsProceedings of the SIGCHI Conference on Human Factors in Computing SystemsProceedings of the 2020 ACM Designing Interactive Systems ConferenceProceedings of the 2021 CHI Conference on Human Factors in Computing SystemsProceedings of the 2018 CHI Conference on Human Factors in Computing SystemsProceedings of the 33rd Annual ACM Conference on Human Factors in Computing SystemsEngineering design, pp. 1–10. External Links: ISBN 9781450367080, Link Cited by: §4.1.
  • A. Thudt, U. Hinrichs, S. Huron, and S. Carpendale (2018) Self-reflection and personal physicalization construction. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, pp. 1–13. External Links: ISBN 9781450356206, Link Cited by: §5.3, §5.3, §5.5.2.
  • B. Ullmer, H. Ishii, and R. J. K. Jacob (2005) Token+constraint systems for tangible interaction with digital information. ACM Trans. Comput.-Hum. Interact. 12 (1), pp. 81–118. External Links: ISSN 1073-0516, Link, Document Cited by: §5.4.
  • A. Vallgårda and J. Redström (2007) Computational composites. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp. 513–522. External Links: ISBN 9781595935939, Link Cited by: §1, §4.2.
  • A. Vande Moere and S. Patel (2010) The physical visualization of information: designing data sculptures in an educational context. In Visual Information Communication, M. L. Huang, Q. V. Nguyen, and K. Zhang (Eds.), Boston, MA, pp. 1–23. External Links: ISBN 978-1-4419-0312-9 Cited by: §1.
  • A. Veldhuis, R. Liang, and T. Bekker (2020) CoDa: collaborative data interpretation through an interactive tangible scatterplot. In Proceedings of the Fourteenth International Conference on Tangible, Embedded, and Embodied Interaction, TEI ’20, New York, NY, USA, pp. 323–336. External Links: Document, ISBN 9781450361071, Link Cited by: §4.1, §5.4.
  • J. Vermeulen, K. Luyten, E. van den Hoven, and K. Coninx (2013) Crossing the bridge over norman’s gulf of execution: revealing feedforward’s true identity. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, New York, NY, USA, pp. 1931–1940. External Links: ISBN 9781450318990, Link Cited by: §1, §3.2.3, §3.2.
  • K. von Ompteda (2019) Data manifestation: merging the human world & global climate change. In 2019 IEEE VIS Arts Program (VISAP), Vol. , pp. 1–8. External Links: Document Cited by: §3.1, §4.1, §4.1, §4.1, §4.2, §5.2, §5.3.
  • W. Walker, H. Sung, C. K. Ong, and F. Casalegno (2017) Exploring spatial meaning with a tangible map. In Proceedings of the 6th ACM International Symposium on Pervasive Displays, PerDis ’17, New York, NY, USA. External Links: Document, ISBN 9781450350457, Link Cited by: §4.1.
  • [111] T. Weinberg Woven climate datascapes. Note: 2021-09-09 Cited by: §6.4.
  • M. Weiser (1991) The computer for the 21 st century. Scientific american 265 (3), pp. 94–105. Cited by: §5.2.
  • M. Wiberg and E. Robles (2010) Computational compositions: aesthetics, materials, and interaction design. International Journal of Design 4 (2), pp. 65–76. Cited by: §1.
  • M. Wiberg (2018) The materiality of interaction: notes on the materials of interaction design. MIT Press, Cambridge, MA, USA. Cited by: §2.1, §3.2.2, Figure 4.
  • W. Willett, Y. Jansen, and P. Dragicevic (2017) Embedded data representations. IEEE Transactions on Visualization and Computer Graphics 23 (1), pp. 461–470. External Links: Document Cited by: §2.2.
  • M. P. Woźniak, J. Dominiak, M. Pieprzowski, P. Ładoński, K. Grudzień, L. Lischke, A. Romanowski, and P. W. Woźniak (2020) Subtletee: augmenting posture awareness for beginner golfers. Proc. ACM Hum.-Comput. Interact. 4 (ISS). External Links: Link, Document Cited by: §2.1, §4.1, §4.1, §4.1, §5.2.
  • P. W. Woźniak, M. Zbytniewska, F. Kiss, and J. Niess (2021) Making sense of complex running metrics using a modified running shoe. External Links: ISBN 9781450380966, Link Cited by: §4.1.
  • P. Wyeth (2008) Understanding engagement with tangible user interfaces. In Proceedings of the 20th Australasian Conference on Computer-Human Interaction: Designing for Habitus and Habitat, OZCHI ’08, New York, NY, USA, pp. 331–334. External Links: Document, ISBN 0980306345, Link Cited by: §3.2.2.
  • J. S. Yi, Y. a. Kang, J. Stasko, and J.A. Jacko (2007) Toward a deeper understanding of the role of interaction in information visualization. IEEE Transactions on Visualization and Computer Graphics 13 (6), pp. 1224–1231. External Links: Document Cited by: §3.2.3, §3.2.3, §3.2, §4.3, §5.4.
  • J. Zhao and A. V. Moere (2008) Embodiment in data sculpture: a model of the physical visualization of information. In Proceedings of the 3rd International Conference on Digital Interactive Media in Entertainment and Arts, DIMEA ’08, New York, NY, USA, pp. 343–350. External Links: Document, ISBN 9781605582481, Link Cited by: §1, §2.2, §3.2.2, §5.3.
  • C. Zheng, P. Gyory, and E. Y. Do (2020) Tangible interfaces with printed paper markers. pp. 909–923. External Links: ISBN 9781450369749, Link Cited by: §6.3.1.
  • J. Zimmerman, J. Forlizzi, and S. Evenson (2007) Research through design as a method for interaction design research in hci. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI ’07, New York, NY, USA, pp. 493–502. External Links: ISBN 9781595935939, Link, Document Cited by: §6.3.1, §6.3.1, §6.3.1.