Nowadays, there is a plethora of digital libraries that provide bibliographic metadata such as the ACM DL111https://dl.acm.org/, Bibsonomy (Hotho et al., 2009), dblp (Ley, 2009), Google Scholar222https://scholar.google.com/, Semantic Scholar333https://www.semanticscholar.org/, SpringerLink444https://link.springer.com/ or ResearchGate555https://www.researchgate.net. These interfaces support search, exploration and sometimes also simple query formulation beyond keyword-based search, but they fail to easily satisfy more complex information needs, e.g. ”Find the papers written by any authors of two different institutions A and B” or ”Find authors who only recently started working on a topic”.
A way to overcome these limitations of search interfaces of digital libraries is usage of specialised tools such as GrapAL (Betts et al., 2019)666As of 2022, the GrapAL web interface is no longer supported., which allows query formulation via the graph-query language Cypher (Francis et al., 2018). The big disadvantage with GrapAL is the prerequisite of all users having to be familiar with that all-purpose query language.
To help users access bibliographic metadata more easily, we extend our domain-specific query language and user interface SchenQL (Kreutz et al., 2021). It focuses on aggregation of data and bibliographic functions, such as RELATED KEYWORDS TO "DLs" or PUBLICATIONS WITH HIGHEST CORERANK METRIC. SchenQL offers simple language components which are easily learned by domain-experts as well as casual users of digital libraries. The queries resemble natural language and the graphical user interface helps in query construction. The two aforementioned information needs could be expressed by the following queries: PUBLICATIONS WRITTEN BY ANY DISTINCT 2 OF [(PERSON WORKS FOR "University of Pisa"), (PERSON WORKS FOR "National Research Council, Italy")] and PERSONS AUTHORED (PUBLICATIONS ABOUT KEYWORD "DLs") AND AUTHORED NO (PUBLICATIONS ABOUT KEYWORD "DLs" WITH YEAR AT MOST 2019).
|query||C — J — K — PU — PE — I — F||query start|
|C||limit? CONFERENCE C?||conference anchor|
|C||(C—WITH DBLPKEY dblpKey—NAMED ? name—ABOUT KEYWORD ((K)—K)—WITH ACRONYM acronym—WITH YEAR COMP year—OF ((PU)—PU))—(WITH ((rank)? (((HIGHEST—LOWEST) METRIC)—(LONGEST—SHORTEST) (name—acronym))—METRIC COMP (h_index—core_rank)—(name—acronym) LENGTH COMP number)) (OPS C)?||conference filters|
|J||limit? JOURNAL J?||journal anchor|
|J||(J—WITH DBLPKEY dblpKey—NAMED ? name—ABOUT KEYWORD ((K)—K)—WITH YEAR COMP year—WITH ACRONYM acronym—WITH VOLUME volume—OF ((PU)—PU))—(WITH ((rank)? (((HIGHEST—LOWEST) METRIC)—(LONGEST—SHORTEST) (name—acronym)) —METRIC COMP (h_index—core_rank)—(name—acronym) LENGTH COMP number)) (OPS J)?||journal filters|
|K||(limit? KEYWORD K?)—(RELATED KEYWORDS TO (((K) (IN ((PU)—PU))?)—K))—((rank)? MOST FREQUENT KEYWORDS OF ((C)—(J)—(PU)—(PE)—C—J—PU—PE))||keyword anchor|
|K||OF ((C)—(J)—(PU)—(PE)—C—J—PU—PE)||keyword filters|
|PU||limit? (PU—(rank)? (MOST CITED—NEWEST—OLDEST) ((PU)—PU))||publication anchor|
|PU||(PUBLICATION—BOOK—ARTICLE—PHDTHESIS—MASTERTHESIS—INPROCEEDING—INCOLLECTION—PROCEEDING) PU?||publication object|
|PU||dblpKey—DOI—? title||publication literals|
|PU||(PU—WITH DBLPKEY dblpKey—WITH DOI doi—WITH ISBN isbn—TITLED ? title—ABOUT (KEYWORD ((K)—K) —TERMS search_terms))—WITH YEAR COMP year—APPEARED IN ((C)—C—(J)—J)—(CITED BY— REFERENCES ((PU)—PU)—(EDITED—WRITTEN) BY ((PE)—PE)—WRITTEN BY ANY DISTINCT? number OF [((PE)—PE)+]—PUBLISHED WITH ((I)—I)))—(WITH (((rank)? (MOST—LEAST) ((REFERENCES (TO ((PU)—PU))?)—(CITATIONS (FROM ((PU)—PU))?)))—((rank)? ((HIGHEST— LOWEST) METRIC)—(LONGEST—SHORTEST) (title—abstract))—(METRIC COMP (h_index—core_rank))—((title—abstract) LENGTH COMP number)—(COMP number ((REFERENCES (TO ((PU)—PU))?)— (CITATIONS (FROM ((PU)—PU))?))))) (OPS PU)?||publication filters|
|PE||limit? (PE—(COAUTHORS OF ((PE)—PE))—((rank)? MOST (PUBLISHING ((PE)—PE) IN ((C)—C—(J)—J))—RESEARCHING ((PE)—PE) ABOUT ((K)—K)))||person anchor|
|PE||(PERSON—AUTHOR—EDITOR) PE?||person object|
|PE||dblpKey—ORCID—(—=)? name||person literals|
|PE||(PE—WITH DBLPKEY dblpKey—NAMED (—=)? name—WITH ORCID orcid—AUTHORED (ONLY—NO)? ((PU)—PU)—EDITED ((PU)—PU)—(CITED BY—REFERENCES) ((PU)—PU)—WORKS FOR ((I)—I)—PUBLISHED IN ((C)—C—(J)—J))—(WITH (((rank)? (MOST—LEAST) ((REFERENCES (TO ((PU)—PU))?)—(CITATIONS (FROM ((PU)—PU))?)))—((rank)? ((HIGHEST—LOWEST) METRIC)—((LONGEST—SHORTEST) name)—((MOST—LEAST) COAUTHORS (IN ((PU)—PU))?))—(METRIC COMP (h_index—core_rank))— (COMP number COAUTHORS (IN ((PU)—PU))?)—(name LENGTH COMP number))) (OPS PE)?||person filters|
|I||limit? (I—((rank)? MOST RESEARCHING ((I)—I) ABOUT ((K)—K)))||institution anchor|
|I||INSTITUTION I?||institution object|
|I||dblpKey—? name||institution literals|
|I||(WITH DBLPKEY dblpKey—NAMED ? name—WITH CITY city—WITH COUNTRY country—WITH MEMBERS ((PE)—PE))—(WITH (((rank)? ((HIGHEST— LOWEST) METRIC)—((LONGEST—SHORTEST) (name—location))— (METRIC COMP (h_index—core_rank))—((name—(location) LENGTH COMP number))) (OPS I)?||institution filters|
|F||(COUNT (query))—(CORE RANKS FOR ((PE)—PE) (IN ((C)—(J)—(PU)—C—J—PU))?)—(limit? ALTERNATIVE NAMES FOR ((C)—(J)—(I)—(PE)—C—J—I—PE)—(limit? MOST FREQUENT attribute OF (query))—(limit? [attribute+] OF (query)))||general functions|
|METRIC||(CORERANK METRIC)—(H-AVG METRIC)||metric|
|COMP||(AT LEAST—AT MOST—MORE THAN—LESS THAN)?||comparator|
|OPS||AND—OR—AND NOT—OR NOT||operators|
SchenQL (Kreutz et al., 2021) describes both the underlying domain-specific query language and the graphical user interface. We extend its previously presented functional range. Our main goal is to support all types of users of digital libraries in their information needs, even domain-experts who tend to use more sophisticated search options compared to casual users (Zavalina and Vassilieva, 2014). Thus, we offer a broad selection of domain-specific functionalities. Our system supports query formulation as well as information exploration but focuses more on the formulation as users of digital libraries have been found to prefer searching to browsing in some domains (Zavalina and Vassilieva, 2014). While our query language can be applied to any digital library, it has been designed with the dblp computer science bibliography777https://dblp.org/ as central use case. Therefore, some language components model the particularities of dblp, such as using keys for referencing publications and persons (e.g., “homepages/f/EdwardAFox”) and using numerical suffixes for disambiguating author names (e.g., “Wei Wang 0042”).
We extend and redefine SchenQL (Kreutz et al., 2021), but all previously described functionality remains (a hidden) part of it. On top of that, we add significant extensions that comprise more and more powerful aggregations, more and more powerful functions (that were previously called filters), and a modified syntax that resembles natural language even more.
Table 1 presents the grammar of SchenQL in a compact form. SchenQL consists of the following 6 base concepts which are interrelated: conferences, journals, keywords, publications, persons and institutions. Queries can be constructed from these base concepts. Here, singular and plural of base concepts are possible (e.g. PERSON and PERSONS). If a base concept is used, entities of this type are retrieved for that part of a query. Occurrences of base concepts in queries can be replaced by specialisations of these base concepts (e.g. BOOK instead of PUBLICATION).
Filters on base concepts restrict the retrieved entities of the target base concept type (e.g. CONFERENCE WITH ACRONYM "JCDL" restricts the result to conferences with the specified acronym or ARTICLE WITH YEAR 2022 restricts the result to publications of type article which appeared in 2022). Filters start with a WITH to help resemble natural language if the respective function describes an attribute that entities of the base concept have. For example, a publication has a year (thus PUBLICATION WITH YEAR 2022), but a publication is about keywords (thus PUBLICATION ABOUT KEYWORD "dsql"). Some filters are aggregations such as ARTICLES WITH HIGHEST CORERANK METRIC or AUTHORS WITH LEAST COAUTHORS. When observing keywords or words, we offer full-text search for words over titles and abstract (e.g. PUBLICATIONS ABOUT TERMS "(digital:libraries)|dsql"), while keywords are a list of topics associated with the publication (e.g. PUBLICATIONS ABOUT KEYWORDS ["digital libraries", "dsql"]).
Filters on base concepts can be concatenated with Boolean operators (e.g. CONFERENCE WITH ACRONYM "JCDL" OR WITH YEAR 2000 or INSTITUTIONS WITH COUNTRY "DE" AND NOT WITH CITY "Cologne"). An AND operator is implicitly assumed between multiple filters and can be dropped (e.g. ARTICLES WITH YEAR AT LEAST 2010 AND WITH YEAR AT MOST 2020 is equivalent to ARTICLES WITH YEAR AT LEAST 2010 WITH YEAR AT MOST 2020).
Literals can be utilised to identify single or multiple entities of a specific base type. They can be used in filters instead of the longer formulation mentioning the base concept specifically (e.g. shortening PUBLICATIONS APPEARED IN (CONFERENCE WITH ACRONYM "JCDL") to PUBLICATIONS APPEARED IN "JCDL" by using the acronym literal).
To enable different types of searches based on string literals like author names, certain prefixes can be used: indicates a search, where all blank-separated specified query components need to be part of the name in any order (e.g. PERSONS NAMED "wang wei" returns persons with names Wei-Bo Wang and Wang Wei Lee), while incorporation of = before person names indicates a required exact match of the name. If no prefix is used for search of person names, names with and without the numerical suffix are retrieved (e.g. PERSONS NAMED ="Wei Wang" returns the one person with the unique name Wei Wang but does not return Wei Wang 0042 while PERSONS NAMED "Wei Wang" returns both of them).
Restriction of the number of returned entities of queries can be achieved by either using LIMIT or RANK in specific positions in the grammar (see Table 1). The first variant specifies an upper limit for the number of results to return (e.g. 5 ARTICLES CITED BY "journals/interactions/Myers98"), the second variant observes ranks of resulting entities such that more than the specified amount of results can be returned if multiple occupy the same rank (e.g. CONFERENCES WITH 5 HIGHEST H-AVG METRIC, which may return more than five results if some top conferences have the same H-AVG metric). Here, an application of the rank restriction is only possible if something is aggregated either by using an aggregation filter or the respective anchor points for keywords, publications, persons or institutions (via e.g. MOST CITED or MOST RESEARCHING).
SchenQL offers a number of general functions on bibliographic metadata that do not return entities of base concepts but other data: COUNT returns the number of entities in the following query, CORE RANKS FOR lists the core ranks and the number of occurrences for the following entities, ALTERNATIVE NAMES FOR returns the list of names for the following entities, MOST FREQUENT attribute OF returns the most frequent value of the specified attribute in the following query and attribute OF returns the values for the attributes of the following query.
The complete overview of SchenQL’s grammar with filters and query components can be found in Table 1. Values of non-numeric variables always need to be encompassed by quotation marks.
The SchenQL GUI is an updated form of the SchenQL GUI (Kreutz et al., 2021). The search interface supports query construction with a suggestion and auto-completion feature (shown in Figure 1). Possible components to continue queries at any given point are suggested below, they can be clicked to appear as part of queries. In the component suggestions, we did not include the specialisations as this would overload the interface. Possible following language components are colour-coded. The colour indicates the type of the component e.g. base concepts (such as CONFERENCES or PERSONS), filters (such as WITH YEAR or ABOUT), literals (such as NUMBER or "STRING"), or restrictions (such as LIMIT or RANK). After running a SchenQL query, a search result is shown. Clicking on result items takes a user to the detail view of the specific type, Figure 2 shows a detail view of a person. Here we include details on the publications, citation information and keywords of their work. Additionally, the Ego Graph (Reitz, 2010) visualisation (in the right part of Figure 2) highlights the most frequent co-authors of a person in the middle. The nodes surrounding the middle symbolise the co-authors, the shorter the distance between the middle and the co-author, the more works have been published together. Another visualisation included in person’s, publication’s, conference’s and journal’s detail views is the BowTie (Khazaei and Hoeber, 2017)
. It can be used to estimate an entity’s position in the scientific landscape, as it visualises the number and age distribution of references and citations (see Figure3 for a visual explanation).
|Information need||SchenQL query formulation|
|Find author by name (Betts et al., 2019)||PERSON NAMED ”Christine Betts”|
|Fuzzy author name search (Betts et al., 2019)||PERSON NAMED ”Wei Wang”|
|Find authors of paper (Betts et al., 2019; Bloehdorn et al., 2007; Zhu, 2017)||PERSONS AUTHORED ”conf/acl/BettsPA19”|
|Find most active author on topic in time frame (Betts et al., 2019)||MOST RESEARCHING (PERSON AUTHORED (PUBLICATIONS WITH YEAR AT LEAST 2000 WITH YEAR AT MOST 2020)) ABOUT KEYWORDS ”digital libraries”|
|Find co-authors (Zhu, 2017; Gómez-Villamor et al., 2008)||COAUTHORS OF ”Adam Jatowt”|
|Find articles of author in venue (Bloehdorn et al., 2007)||ARTICLES WRITTEN BY ”Waleed Ammar” APPEARED IN ”JCDL”|
|Find papers citing author (Zhu, 2017)||PUBLICATIONS REFERENCES (PUBLICATIONS WRITTEN BY ”Yongjun Zhu” )|
|Find papers on topics (Betts et al., 2019; Bloehdorn et al., 2007; Zhu, 2017)||PUBLICATIONS ABOUT KEYWORD [”digital libraries”, ”search”]|
|Find papers using keywords (Betts et al., 2019)||PUBLICATIONS ABOUT TERMS ”digital:libraries—dsql”|
|Find papers on topic cited by author’s papers in venue (Zhu, 2017)||PUBLICATIONS ABOUT KEYWORD ”search” CITED BY (PUBLICATIONS WRITTEN BY ”Joeran Beel” APPEARED IN ”JCDL”)|
|Find number of papers on topic a referencing papers on topic b (Betts et al., 2019)||COUNT (PUBLICATION ABOUT KEYWORD ”digital library” REFERENCES (PUBLICATIONS ABOUT KEYWORD ”search”)|
|Find institution of authors of paper (Zhu, 2017)||INSTITUTIONS WITH MEMBERS (PERSONS AUTHORED ”conf/cikm/ZhuSRH08”)|
|Find topics of author’s papers in venue (Zhu, 2017)||KEYWORDS OF (PUBLICATIONS APPEARED IN ”JCDL” WRITTEN BY ”Michael Ley”)|
|Find topics of papers cited by papers on topic a (Zhu, 2017)||KEYWORDS OF (PUBLICATIONS CITED BY (PUBLICATIONS ABOUT KEYWORD ”dsql”))|
|Find persons with 5 longest names||PERSONS WITH 5 LONGEST NAME|
|Find authors of 5 earliest papers on topic||PERSONS AUTHORED 5 OLDEST (PUBLICATIONS ABOUT KEYWORD ”digital library”))|
|Find authors who only recently started working on topic||PERSONS AUTHORED (PUBLICATIONS ABOUT ”digital libraries”) AND AUTHORED NO (PUBLICATIONS ABOUT KEYWORD ”digital libraries” WITH YEAR AT MOST 2017)|
|Find authors who published the most on topic since year (Betts et al., 2019)||MOST RESEARCHING (PERSONS AUTHORED (PUBLICATIONS WITH YEAR AT LEAST 2015)) ABOUT KEYWORD ”dsql”|
|Find author co-citation (Zhu, 2017)||PUBLICATIONS REFERENCES (PUBLICATIONS WRITTEN BY ”Ralf Schenkel”) REFERENCES (PUBLICATIONS WRITTEN BY ”Norbert Fuhr”)|
|Find papers written by members of institutions a and b||PUBLICATIONS WRITTEN BY ANY DISTINCT 2 OF [(PERSON WORKS FOR ”University of Pisa”), (PERSON WORKS FOR ”National Research Council, Italy”)]|
|Find papers from year cited at least 20 times||PUBLICATIONS WITH YEAR 2020 WITH AT LEAST 20 CITATIONS|
|Find papers close to area (Gómez-Villamor et al., 2008)||PUBLICATIONS ABOUT ( 5 MOST FREQUENT KEYWORDS OF (RELATED KEYWORDS TO ”digital libraries”))|
|Find institution in country with highest h index||INSTITUTIONS WITH COUNTRY ”DE” WITH HIGHEST H-AVG METRIC|
|Find topics of papers in conference in year||KEYWORDS OF (PUBLICATIONS WITH YEAR 2018 APPEARED IN ”JCDL”)|
|Find co-words (Zhu, 2017)||RELATED KEYWORDS TO ”digital libraries”|
|Find most cited paper in journal||MOST CITED (PUBLICATION APPEARED IN ”JODL”)|
|Find bibliographic couplings (Zhu, 2017)||PUBLICATIONS CITED BY ”conf/aaai/ChangRRR08” CITED BY ”conf/aaai/HashemiH18”|
|Find co-cited papers (Zhu, 2017)||PUBLICATIONS REFERENCES ”journals/annals/Grudin05” REFERENCES ”journals/interactions/Myers98”|
2.3. System Architecture and Data
The SchenQL front-end uses Node.js 16.13 and incorporates Boot and Spring Boot. It communicates with a REST API which accepts user input queries and relays them to the compiler module. Here, for the lexer and parser, we extended and modified the previous SchenQL grammar which was generated with ANTLR to also support aggregation functions as well as more filters in SchenQL. The SchenQL queries are compiled via Java 17 into an intermediate parse tree structure, before they are translated to an exchangeable target language. We currently run queries against a PostgreSQL888Compared to the previous version of SchenQL (Kreutz et al., 2021), here we decided to use PostgreSQL instead of MySQL for increased performance. 14.1 database. Results of queries are transmitted back through the API to the front-end where they are used to populate the result tables and detail views of the GUI (see Figure 4).
As data source for our system, we combine the dblp XML dump from 1st October 2021999https://dblp.org/xml/release/dblp-2021-10-01.xml.gz with Semantic Scholar data (Ammar et al., 2018) from October 2021 for citation information as well as AMiner Open Academic Graph 2.1 (Sinha et al., 2015; Tang et al., 2008; Zhang et al., 2019) data for identifying automatically generated keywords of publications; abstracts are taken from both collections.
2.4. Application and Information Needs
SchenQL can be used to satisfy typical information needs of users of digital libraries or more complex ones which cannot directly be answered by current digital libraries.
Table 2 contains a selection of typical information needs (upper part) and ones which cannot be directly satisfied by using digital libraries (lower part) as well as their exemplary formulations in SchenQL. The listed information needs are partially mentioned in previous works (Zhu, 2017; Betts et al., 2019; Bloehdorn et al., 2007; Gómez-Villamor et al., 2008). Further information needs such as query refinement, backwards and forwards citation analysis, quality judgement (Hoeber and Khazaei, 2015) as well as support in query formulation and general information exploration are mainly supported by the GUI of SchenQL.
We presented an extension of the SchenQL query language and GUI that incorporates aggregation functions and supports more general functions. Our system aims to support casual users as well as domain-experts by providing a huge selection of domain-specific functionalities. We make these functions accessible especially for casual users by suggesting possible following query components in our GUI.
SchenQL could be extended by enabling the support of graph-based operations (as seen with GrapAL (Betts et al., 2019)) such as the calculation of centrality scores, PageRank or hubs and authorities on a user-defined portion of the bibliographic network. Another line of work could focus on the incorporation of different visualisations to support users’ needs, for example colour-coded topical distributions in publications to help find relevant papers or estimate an author’s fit as a reviewer for a manuscript.
Evaluations of SchenQL could include information search where users are asked to retrieve information for predefined queries such as the ones we presented in Table 2 or more open, task-based studies where user behaviour and preferences could be observed more in detail as seen in other evaluations (Hoeber and Khazaei, 2015). For example sense-making studies where users are asked to find and collect information (Pirolli, 2009) such as the question which one of two researchers could better fit a fictional open academic position, or exploratory search which incorporate information lookup, learning and investigation (Pirolli, 2009) such as asking the users to get familiar with new subject areas could provide interesting insights and highlight further directions.
- Ammar et al. (2018) Waleed Ammar, Dirk Groeneveld, Chandra Bhagavatula, Iz Beltagy, Miles Crawford, Doug Downey, Jason Dunkelberger, Ahmed Elgohary, Sergey Feldman, Vu Ha, Rodney Kinney, Sebastian Kohlmeier, Kyle Lo, Tyler Murray, Hsu-Han Ooi, Matthew E. Peters, Joanna Power, Sam Skjonsberg, Lucy Lu Wang, Chris Wilhelm, Zheng Yuan, Madeleine van Zuylen, and Oren Etzioni. 2018. Construction of the Literature Graph in Semantic Scholar. In NAACL-HLT (3). Association for Computational Linguistics, 84–91.
- Betts et al. (2019) Christine Betts, Joanna Power, and Waleed Ammar. 2019. GrapAL: Connecting the Dots in Scientific Literature. In Proceedings of the 57th Conference of the Association for Computational Linguistics, ACL 2019, Florence, Italy, July 28 - August 2, 2019, Volume 3: System Demonstrations, Marta R. Costa-jussà and Enrique Alfonseca (Eds.). Association for Computational Linguistics, 147–152. https://doi.org/10.18653/v1/p19-3025
- Bloehdorn et al. (2007) Stephan Bloehdorn, Philipp Cimiano, Alistair Duke, Peter Haase, Jörg Heizmann, Ian Thurlow, and Johanna Völker. 2007. Ontology-Based Question Answering for Digital Libraries. In Research and Advanced Technology for Digital Libraries, 11th European Conference, ECDL 2007, Budapest, Hungary, September 16-21, 2007, Proceedings (Lecture Notes in Computer Science, Vol. 4675), László Kovács, Norbert Fuhr, and Carlo Meghini (Eds.). Springer, 14–25. https://doi.org/10.1007/978-3-540-74851-9_2
- Francis et al. (2018) Nadime Francis, Alastair Green, Paolo Guagliardo, Leonid Libkin, Tobias Lindaaker, Victor Marsault, Stefan Plantikow, Mats Rydberg, Petra Selmer, and Andrés Taylor. 2018. Cypher: An Evolving Query Language for Property Graphs. In Proceedings of the 2018 International Conference on Management of Data, SIGMOD Conference 2018, Houston, TX, USA, June 10-15, 2018, Gautam Das, Christopher M. Jermaine, and Philip A. Bernstein (Eds.). ACM, 1433–1445. https://doi.org/10.1145/3183713.3190657
- Gómez-Villamor et al. (2008) Sergio Gómez-Villamor, Gerard Soldevila-Miranda, Aleix Giménez-Vañó, Norbert Martínez-Bazan, Victor Muntés-Mulero, and Josep Lluís Larriba-Pey. 2008. BIBEX: a bibliographic exploration tool based on the DEX graph query engine. In EDBT 2008, 11th International Conference on Extending Database Technology, Nantes, France, March 25-29, 2008, Proceedings (ACM International Conference Proceeding Series, Vol. 261), Alfons Kemper, Patrick Valduriez, Noureddine Mouaddib, Jens Teubner, Mokrane Bouzeghoub, Volker Markl, Laurent Amsaleg, and Ioana Manolescu (Eds.). ACM, 735–739. https://doi.org/10.1145/1353343.1353439
- Hoeber and Khazaei (2015) Orland Hoeber and Taraneh Khazaei. 2015. Evaluating citation visualization and exploration methods for supporting academic search tasks. Online Inf. Rev. 39, 2 (2015), 229–254. https://doi.org/10.1108/OIR-10-2014-0259
- Hotho et al. (2009) Andreas Hotho, Robert Jäschke, Dominik Benz, Miranda Grahl, Beate Krause, Christoph Schmitz, and Gerd Stumme. 2009. Social Bookmarking am Beispiel BibSonomy. In Social Semantic Web: Web 2.0 - Was nun?, Andreas Blumauer and Tassilo Pellegrini (Eds.). Springer, 363–391. https://doi.org/10.1007/978-3-540-72216-8_18
- Khazaei and Hoeber (2017) Taraneh Khazaei and Orland Hoeber. 2017. Supporting academic search tasks through citation visualization and exploration. Int. J. Digit. Libr. 18, 1 (2017), 59–72. https://doi.org/10.1007/s00799-016-0170-x
- Kreutz et al. (2021) Christin Katharina Kreutz, Michael Wolz, Jascha Knack, Benjamin Weyers, and Ralf Schenkel. 2021. SchenQL: in-depth analysis of a query language for bibliographic metadata. Int J Digit Libr (2021). https://doi.org/10.1007/s00799-021-00317-8
- Ley (2009) Michael Ley. 2009. DBLP - Some Lessons Learned. Proc. VLDB Endow. 2, 2 (2009), 1493–1500. https://doi.org/10.14778/1687553.1687577
- Pirolli (2009) Peter Pirolli. 2009. Powers of 10: Modeling Complex Information-Seeking Systems at Multiple Scales. Computer 42, 3 (2009), 33–40. https://doi.org/10.1109/MC.2009.94
- Reitz (2010) Florian Reitz. 2010. A Framework for an Ego-centered and Time-aware Visualization of Relations in Arbitrary Data Repositories. CoRR abs/1009.5183 (2010). arXiv:1009.5183 http://arxiv.org/abs/1009.5183
- Sinha et al. (2015) Arnab Sinha, Zhihong Shen, Yang Song, Hao Ma, Darrin Eide, Bo-June Paul Hsu, and Kuansan Wang. 2015. An Overview of Microsoft Academic Service (MAS) and Applications. In WWW (Companion Volume). ACM, 243–246.
- Tang et al. (2008) Jie Tang, Jing Zhang, Limin Yao, Juanzi Li, Li Zhang, and Zhong Su. 2008. ArnetMiner: extraction and mining of academic social networks. In KDD. ACM, 990–998.
- Zavalina and Vassilieva (2014) Oksana Zavalina and Elena Vassilieva. 2014. Understanding the Information Needs of Large-Scale Digital Library Users. Library Resources & Technical Services 58 (04 2014), 84. https://doi.org/10.5860/lrts.58n2.84
- Zhang et al. (2019) Fanjin Zhang, Xiao Liu, Jie Tang, Yuxiao Dong, Peiran Yao, Jie Zhang, Xiaotao Gu, Yan Wang, Bin Shao, Rui Li, and Kuansan Wang. 2019. OAG: Toward Linking Large-scale Heterogeneous Entity Graphs. In KDD. ACM, 2585–2595.
- Zhu (2017) Yongjun Zhu. 2017. Graph-based Interactive Bibliographic Information Retrieval Systems.