Analyzing the Impact of Covid-19 Control Policies on Campus Occupancy and Mobility via Passive WiFi Sensing

by   Camellia Zakaria, et al.

The outbreak of has resulted in many different policies being enacted across the world to reduce the spread of the virus. These efforts range from increased social awareness and social distancing (Sweden) to full country-wide lockdowns (Singapore, most of Europe, many states in the US). The effectiveness of various policies in containing the spread of the disease is still being studied by researchers. In this paper, using WiFi data collected directly from the infrastructure, we present a detailed analysis of the impact of related policies on the staff and students of three different college campuses. Two of these campuses are in Singapore, while the third is in the Northeastern United States. Our study focuses on two different key metrics, 1) Occupancy, defined as the number of people in a building, on average, in a specific hour, and 2) Mobility, defined as the number of places visited by an individual, on average, in a specific hour. We use these two metrics since they provide a strong indicator of how likely it is for to spread if there is an outbreak. Our results show that online learning, split-team, and other space management policies are effective at lowering Occupancy. However, they do not change the Mobility patterns for individuals who are still moving around. Reducing Mobility requires introducing strict stay-at-home or lockdown orders, but doing so increases Occupancy in residential spaces. We present our results and then discuss the implications of the findings for policymakers.



There are no comments yet.


page 6

page 7

page 8


Influence of Mobility Restrictions on Transmission of COVID-19 in the state of Maryland – the USA

Background: The novel coronavirus, COVID-19, was first detected in the U...

Data-Driven Modeling Reveals the Impact of Stay-at-Home Orders on Human Mobility during the COVID-19 Pandemic in the U.S

If the novel coronavirus (COVID-19) is to reduce human travel by imposin...

Multimodal Indoor Localisation for Measuring Mobility in Parkinson's Disease using Transformers

Parkinson's disease (PD) is a slowly progressive debilitating neurodegen...

Quarantine Fatigue: first-ever decrease in social distancing measures after the COVID-19 pandemic outbreak before reopening United States

By the emergence of the novel coronavirus disease (COVID-19) in Wuhan, C...

Current Implicit Policies May Not Eradicate COVID-19

Successful predictive modeling of epidemics requires an understanding of...

Understanding individual behaviour: from virtual to physical patterns

As "Big Data" has become pervasive, an increasing amount of research has...
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

The COVID-19 epidemic became a global pandemic because of the highly contagious nature of the virus. Countries that rapidly enacted early case identification and quarantine of infected persons have been more successful at containing the spread of infection and reducing deaths (Organization, 2020b). Singapore was initially very successful at controlling the spread of COVID-19 , by following an incremental approach to social distancing involving a variety of social controls such as splitting work teams and having returning residents stay at home (Yong, 2020). This strategy allowed the economy to continue running while still providing effective COVID-19 control measures.

Unfortunately, the measured approach initially taken in Singapore was found inadequate when COVID-19 began spreading prolifically in densely packed manual worker dormitories (Ministry of Health, 2020d, b). This phenomenon resulted in a revision of policy and a country-wide lockdown that started in April 2020. Note, the lockdown is still in force at the time of writing of this paper (Ministry of Health, 2020c; Ministry of Education, 2020a).

A different approach to the pandemic has taken by countries like the United States of America and Brazil. These countries did not implement national policies to contain COVID-19, and instead left infection control to the discretion of state/regional governors. This distributed approach has generally resulted in poorer outcomes and higher death rates compared to countries who have taken consistent country-wide actions (Organization, 2020a, b).

In addition to enacting policy, it is beneficial to evaluate the effectiveness of different movement controls as this could facilitate more principled planning for gradual re-opening of locked down societies.

In this paper, we present findings from production WiFi networks, that show the effect of various COVID-19 related policies on the measures of occupancy and mobility of students and staff across three different university campuses. Two of these campuses are in Singapore, and one is in the northeastern part of the United States of America.

We define Occupancy as the number of people in each building, on average, in a specific hour (e.g. 12 p.m. - 1 p.m. lunchtime). It also refers to as the occupancy density or population density of a building. Mobility is defined as the number of places visited by an individual, on average, in a specific hour.

We use these metrics as they give a reliable indicator of how likely it is for COVID-19 to spread if there is an outbreak. Occupancy suggests how likely it is for COVID-19 to spread within a residence or building as higher occupancy places are more crowded and thus more likely to be affected if an outbreak occurs. Mobility shows how likely it is for a COVID-19 

outbreak to spread beyond the infected individual’s home and work environment. In particular, if an infected individual has a high mobility or is close to other individuals with high mobility, an outbreak will likely spread to all those places that are visited by the high mobility individuals. Indeed, the movement of individuals either visiting a country or returning home was the primary vector through which

COVID-19 spread throughout the world.

For both these metrics, we expect to see the time of the day (morning versus lunch versus dinner, etc.) and day of the week (weekday versus weekend) effects for both of these metrics. The goal of this paper is to show how various COVID-19 related policies change the normally expected patterns for these two metrics. Our results show that even gradual policies have an effect on occupancy but minimal impact on mobility and that only strict quarantine measures are effective at limiting mobility across campus. In particular, this paper contributes to the following:

  • It provides one of the first detailed looks at the effect of COVID-19 related policies on three campus populations across two different countries. We provide detailed results showing how the occupancy and mobility metrics change across all three campuses as COVID-19 related quarantine policies, of increasing severity, are enacted.

  • It shows that controlling the mobility of people across campus requires strict quarantine policies.

  • It provides evidence that passive monitoring of campus occupancy and mobility patterns can be of great help to administrators deciding on the appropriate policies to combat future pandemics and other contagious outbreaks.

2. Data Collection

Our data sets were collected directly from the production WiFi networks of three different university campuses. Two of these campuses ( and ) are in Singapore, and the last one () is in the North-Eastern portion of the United States of America. Two campuses are full-sized residential campuses with over 200 buildings each and 40K to 50K students and staff while the last university is a small boutique non-residential establishment with 10,000 students and staff spread across a handful of buildings.

All three universities run Aruba-equipment supplied WiFi networks with one university also running a Cisco-equipment supplied WiFi network in addition to an Aruba network. For the Aruba networks, we pulled the WiFi data directly from the infrastructure using either real-time location services (RTLS) APIs (Networks, 2020) or by reading the system logs directly. For the Cisco network, we pulled WiFi data directly from the network using the Cisco Connected Devices (CMX) Location API v3 (Systems, 2020) (recently rebranded as Cisco DNA Spaces). In all cases, we obtained the following information: for all associated WiFi devices, the timestamp when the associated device was seen, the BSSID of the Access Point (AP) that saw the device, and the hashed client MAC address of the associated device. For two of the campuses, we can also obtain the same information (time seen, BSSID that saw the device, and hashed client MAC) for unassociated devices as well – i.e. devices with WiFi on that are just scanning.

For all three campuses, we have associated device data from Feb 2020 onwards. This allows us to have a very clear view of campus occupancy and mobility patterns across campus both before, during, and after COVID-19 related measures were implemented. Table 1 provides details of each of the 3 campuses as well as the data collected.

Campus Campus Name No. Buildings No. Students No. Staff No. APs
Singapore Management University(SMU) 7 9,000 1,000 800
National University of Singapore(NUS) 240 40,000 10,000 13,000
University of Massachusetts, Amherst 230 30,000 8,000 5,500
Table 1. Details of Each Campus Studied in this Paper

2.1. Ethical Considerations

All data used in this paper was obtained directly from the campus infrastructure and bounded by the computing agreements agreed to by each WiFi user when they received their WiFi credentials. These agreements allowed us to use their data for aggregate analysis as long as individual identifiers were not used. As such, every MAC address obtained from the WiFi infrastructure was hashed using a consistent 1-way hash function, and no specific user details (e.g., login IDs) were used. All of the analysis used by this paper focuses on large aggregates with no analysis of specific individuals performed.

3. Mobility and Occupancy Analysis for Singapore Universities

We aimed to investigate the adequacy of safety measures to curb the spread of COVID-19 enforced by the respective university administrations on their staff and students. We did this solely through analysing mobility data generated from the WiFi data collected directly from the WiFi infrastructure, as explained in Section 2.

We used a three-step spatio-temporal analysis by first, generating unique device counts connected to every access point (AP) on campus, and applying heuristics to create user groups for each device. Second, the WiFi data was processed to generate transition records of each device that listed the locations (with the dwell duration) each device had visited for a given day. Third, and finally, we examined the changes in space density and movements on-campus to determine the effectiveness of various safety measures put in place at significant times points of

COVID-19. These time points are summarised in Table 2.

Singapore Phases Date Started Safety Measures
Pre COVID-19 Before Awareness on personal hygiene
() 19 Feb, 2020  moves some classes online
Phase 1 () 19 Feb, 2020 National threat level raised to orange (Ministry of Health, 2020e)
14-days home quarantine enforcing if returning from China (Ministry of Manpower, 2020)
All core curriculum moved to online learning for
In-class mid-term assessment cancelled for
both  and  implement a 1 meter safe distancing policy
 closes all sports facilities
Classes ¿= 50 students moved to online learning for  and
Phase 2 () 22 Mar, 2020 All travel cancelled unless mandatory
All visitors to Singapore issued 14 day Stay Home order (Immigation & Checkpoints Authority, 2020)
 enforcing A/B shifts where all students & staff
must alternate being offsite every other week
Phase 3 () 3 Apr, 2020 Full shift to online learning for all schools at all levels (Ministry of Education, 2020b)
All exams moved online at  and
Pass / Fail option offered to students at  and
 only allowing key personal on campus
 allowing most employers to work from home
Phase 4 7th April 2020 Full country-wide stay at home quarantine (Ministry of Health, 2020c)
(, ongoing) Nobody allowed on campus for . All buildings closed.
Only approved students allowed to stay in dorms at
Approved students allowed to travel to  to study
Measures extended until Jun 2020 (Ministry of Education, 2020a)
Table 2. Description of the five phases of COVID-19 related safety measure enacted in Singapore and at  and  

3.1. Overall Control Policy

Singapore was first alerted of ’severe pneumonia’ cases in Wuhan city on 2nd January 2020. From that point on, Singapore’s Ministry of Health (MOH) has mandated a series of escalating control policies to prevent high infection rates of COVID-19 while minimizing significant disruptions to the daily routines of its residents. The first case of COVID-19 in Singapore was reported on 22nd January, and more cases started appearing over the next few days. As shown in the subsequent phases ( to ), Singapore’s MOH took increasingly strong measures to contain the spread of infection. These measures included mandatory stay-at-home quarantine orders for visitors, moving all academic programs online, to enforcing country-wide stay-at-home orders. Additionally, numerous facilities across Singapore were re-purposed as quarantine centres. For this analysis, several dormitory blocks at the  campus were re-purposed for this use in early May 2020 (redacted, 2020).

3.2. Impact of Different Policies on Occupancy

We first computed the drop in campus occupancy in Singapore as each phase of COVID-19 related policies were enacted. Figure 1 plots the unique device counts for one building within the Singapore universities,  and , over the COVID-19 phases, to . We observed there was a more than 90% percentage drop from phase to and beyond for  when the university implemented an almost full work-from-home policy () followed by the nationwide lockdown (). Despite taking the same set of measures,  was successful in reducing occupancy to only 75% at , and 98% by the lockdown. For , the drop was almost 100% by Phase 4 () as nobody, except for security staff, was allowed onto campus whereas  still allowed a few thousand students to stay in the dorms.

Figure 1. Space occupancy for one school building for each university,  (blue) and  (orange), plotted from to .
Figure 2. The percentage change in occupancy for a recreational area, library and food court within  (left) and  (right), plotted from pre COVID-19 time () to current time ().

To understand this change of occupancy in more detail, Figure 2 shows the percentage change in space occupancy for different types of areas – in particular, areas used for recreation (e.g. indoor gym, outdoor spaces), studying (e.g. library), and dining (e.g. food courts). The values for  are shown in the left figure, while the values for  are shown in the right figure. The absolute count for each percentage is listed inside each area (e.g. The absolute count for _Food for  in is 22 people comprising about 40% of the total occupancy).

There were significant differences in the space utilisation between the two campuses. For example,  closed all recreational facilities in onwards, and this is reflected in the significant percentage drop. The on-campus dining facilities were also mostly closed from onwards.

As  has a large number of students staying on campus in dorms ( does not have dorms on campus), even in , the occupancy of recreational spaces was relatively high (and higher percentage-wise than earlier phases). This result is likely because students staying in dorms did not want to stay in their rooms all day long and decided to go out to these recreational spaces (which is technically a violation of the quarantine rules in effect).

Such occurrences raise concerns that these recreational spaces would have larger than optimal occupancy densities and undesirable mixing students from different dorms that would compromise measures designed to contain the spread of infection.

3.3. Impact of Various Policies on Mobility

Figure 3. Device transitions originating from _building1 to five other buildings. The patterns of mobility remained the same on the overall despite significant reduction in the number of transitions over time.
Figure 4. Device transitions originating from _building1 to three other academic buildings. The mobility rate is significantly less on the overall.

Next, we sought to determine how various social mobility control measures influenced mobility patterns across both universities. First, looking at , Figure 3 charts the average transitions made on a per-building level over a day in each phase, from one building (called _building1), to five other  buildings. The transition count indicates the number of times a person moved from _building1 to the indicated building on that day.

The results showed an expected decrease in the overall numbers as each phase was introduced, beginning with the university’s decision to conduct online learning for its students at phase . In phase ,  introduced full A/B schedules where only half the student and staff population could physically be on campus at any one time. This step reduced the overall occupancy (as shown in Figure 1 and decreasing Transition count for in Figure 3).

However, the actual mobility rate has not decreased as, even though the total number of people on campus had decreased, the number of buildings that each person visited had not. This is understandable as the work that each person on campus had to do remained the same – requiring them to visit the same set of buildings that they had to previously.  reduced the campus occupancy to almost 0 in Phases onwards, and this naturally reduced the mobility rate.

Figure 5. The hourly trend of occupancy for  plotted before () and during full lockdown ().

We next looked at data from  to understand if these changes in mobility patterns were consistent. Figure 4 shows the absolute number of transitions (numbers within each bar) along with the percentage of transitions from one  academic building to three other academic buildings. Similar to , even though the total occupancy of the campus decreased due to the measures enacted in , the mobility rate (amongst those staff and students still on campus) remained high until more complete lockdown policies were enacted in onwards.

Figure 6. Transitions of 200 sampled  occupants to different (neighbourhood) area (e.g., another dorm) and unique locations within the vicinity of .

3.4. On-campus Living

The previous analysis focused mostly on academic buildings. However, , unlike , has a significant student population still residing in on-campus dormitories even during the full lockdown phase . We decided to dig deeper and understand the behaviour of these students living in these  dorms.

Figure 5 plots the daily activity trend of one dormitory location (called  ) over phases and . Overall, we observed the same daily occupancy levels (approximately 700) for this dorm across all phases – indicating that the population of the dorm had not changed significantly between phases.

However, our analysis of mobility patterns amongst students staying in the dorms revealed some interesting findings. In particular, we found that even during lockdown periods (), a significant number of people were moving actively across campus – which is technically a violation of the lockdown rules.

Figure 6 presents the mobility analysis of 200 randomly selected individuals staying in . We found that their mobility rate increased during and compared to . Each occupant was making at least three transitions on average in phase , which doubled to about six transitions in onwards.

While the numbers seem surprisingly high at first glance, we discovered that the increases in mobility were due to individuals moving between buildings in the same dorm (to visit the dining and recreational facilities) and visiting bus stops more often (to head to other dining and grocery areas). These were all shorter length transitions compared to earlier phases – in earlier phases, the transitions had long durations as the individuals were going to academic buildings for coursework reasons.

3.5. Main Takeaways

From the analysis of  and , the main takeaway we derive is that policies that allow telecommuting and split-team load balancing are excellent for reducing the people density on campus. However, the staff and students that do work on campus are more likely to continue visiting the same set of places they utilized, thus, can lead to serious issues if an outbreak occurs – as the COVID-19 can be easily spread to all the other people visiting those areas. Thus, to avoid uncontrolled outbreaks, it may be necessary to limit the mobility of individuals and the only policy that was successful at this (from the many policies that were tried in Singapore) is a full lockdown where everyone is given stay-at-home orders.

In addition, the mobility analysis of dorm occupants at  suggests that reducing mobility will require providing everything occupants need at their premises itself. Otherwise, the mobility rate could go up (even if the actual length of the transitions are shorter in duration) as individuals travel to other places to procure food and other essential items needed during a lockdown.

4. Mobility and Occupancy Analysis for US University

Unlike Singapore, the US state our campus () was located in only had a single response – the state went from business as usual to a stay-at-home order with shutdowns of many businesses (Lee, 2020a)111Note: for anonymity reasons, we cite an article listing all the states that have effected a similar policy. .

We observed similar occupancy trends, compared to  and , as  transitioned into a lockdown. Figure 7 plots the total number of unique devices detected for different area types over ten days for each of the two phases. The areas picked were “Recreational” (e.g. gym), Dorm (e.g. on-campus dormitory housing), “Lib” (campus libraries), and “Food” (e.g. food courts). Overall, we observed a more than 90% decrease in occupancy between phases (business as usual) (full stay-at-home orders). In addition, we observed that the quarantine policies had significantly shifted the occupancy rates with the dormitories becoming the most occupied locations during and the occupancy of the previously busy library areas reducing significantly.

US State Phases Date Started Safety Measures
Pre COVID-19 () 29 Feb, 2020 No Policy. Business as Usual.
Phase 1 () 20 Mar, 2020 Stay at Home State Wide Order (Lee, 2020a)
No classes at
Dorms cleared at  except in special cases
Table 3. Dates corresponding to the safety measures for COVID-19 in
Figure 7. Space occupancy within  for both and – broken down by usage type
Figure 8. CDF of unique transition locations made by occupants of , plotted from pre COVID-19 time ( the blue line on the right) to the current quarantine phase ( the red line on the left)

We next investigated if similar changes to the mobility rate had occurred due to the quarantine policy. Figure 8 plot the CDF of the number of other places visited by the occupants of one particular  dorm (called ) in each phase. We observed that the number of visits had significantly reduced with the percentile reducing by slightly more than half (about ten visits in versus less than five visits in ) and the percentile decreasing from about 17 visits to about nine visits.

Figure 9. The percentage change in transitions for different areas made by occupants of , plotted from pre COVID-19 time () to current time ().

Figure 9 breaks down these visits by the type of place visited. We observed that most of the visits in were to dining and recreational locations. However, in , most of the visits were made to another on-campus dorm. We believe this was to avail of the dining and recreational facilities available at that dorm as certain previously-popular places on campus had been closed in .

5. Discussion of Policy Impact and Limitations of Our Study

In this section, we discuss the impact of the policy decisions on the probabilities of

COVID-19 spreading amongst the student population. In particular, we looked at two different modes by which COVID-19 could spread: 1) spreading amongst the people who are in the same place as where you stay. We call this vector as local spread. 2) spreading amongst people that you meet as you travel to and spend time at places other than your primary place of residence. We call this vector as mobility spread. Controlling each of these vectors requires different approaches.

In particular, controlling local spread requires reducing the density of people that you stay with while controlling mobility spread requires reducing the amount of movement outside their homes that individuals perform. Note: both these approaches also apply in both cases just with different intensities. For example, limiting the movement you do at home can control local spread but this may not practical in many homes while reducing the density of people, in general, can reduce mobility spread as well.

5.1. Controlling Local Spread Across Campus

From the results presented in Sections 3 and 4, we note that quarantine policies were very effective in removing people from their workplaces. The policies immediately removed one primary source of local spread – i.e., spreading a virus amongst your co-workers.

However, this quarantine policy resulted in higher densities being observed in the student dormitories, as shown in Figure 7 where the occupancy in dorms increased after the initial quarantine measures were imposed. This was “solved” by the universities asking students to vacate the dorms and return home. This reduced the density of the dormitories as shown by our results, but it moved the problem elsewhere. In particular, we believe that many home residences would have seen much higher population densities as a result of these quarantine policies as the entire country (Singapore) and state (in the US) was asked to stay at home for extended periods of time. This could make it easier for residents to fall sick if someone in their vicinity had the virus.

Indeed, Singapore experienced this first-hand as the second wave of COVID-19 outbreak in Singapore occurred in the dormitories used by foreign workers. The population densities at these dormitories were very high, and the first cases of COVID-19 were reported on April 1st 2020 (Ministry of Health, 2020d). This spread grew very fast and resulted in thousands of infections within the dormitories within just a few weeks (Ministry of Health, 2020b). Fortunately, the strict quarantine policies, enacted just a few days (on 7th April 2020) after the first dorm infection (on 1st April 2020) when the authorities realised the potential for the spread to grow out of control, ensured that the virus was contained within the dormitories. For example, while the infection rate remains high in the worker dormitories, the number of cases in the rest of Singapore is almost non-existent – on May 15th 2020, 791 new infection cases were discovered in the worker dorms with just one other case discovered across the rest of Singapore (Ministry of Health, 2020a). Thus, while quarantine policies can lead to higher local transmission rates, the significantly reduced mobility stops the virus from spreading beyond the local area.

5.2. Controlling Mobility Spread

The second mode of virus propagation is where an infected individual travels to another place and infects someone there. We call this mobility spread. This vector is particularly dangerous as it can allow the virus to spread to formally safe areas very quickly. Indeed, it was this vector that was responsible for spreading COVID-19 throughout the world – carried by infected individuals travelling between countries.

We see from the results in Section 3 that only a strict quarantine was effective in reducing mobility patterns. In particular, the split team and other approaches used in phases and did not have a significant impact on the mobility patterns of individuals (defined as the number of unique places visited by an individual in a day). However, when strict quarantine policies were enacted, starting in and fully enacted in , we note that the amount of individual mobility has significantly reduced. Also, when people were mobile, they spent significantly lower amounts of time at each place visited.

This data backs up the policy decisions in both Singapore and the US state to enact a strict quarantine as the impact on individual mobility is very clear. Such measure, in turn, dramatically reduces the probability that COVID-19 can spread beyond a local area. However, as stated previously, reducing mobility comes at the expense of increasing the population density of homes, dormitories, and other residential areas. Hence, there could be a higher probability of local infections as a result of a strict quarantine (as demonstrated by the worker dormitories outbreak in Singapore).

5.3. Scaling to Other Areas

In this paper, we showed how to use WiFi data collected directly from the infrastructure to provide a detailed view of the effectiveness of various COVID-19 related containment policies. We believe this is a very viable approach for other locations as it a) does not require deploying an app that needs to be installed (the actual usage rate of the various over-hyped Bluetooth-based contact tracing apps is very low at 5% or lower (Aravindan and Phartiyal, 2020; Lee, 2020b)), b) can work anywhere with WiFi or even if cellular data is available, and c) scales very easily to the number of devices that have that interface (cellular or WiFi) enabled. The analysis can also provide privacy by leveraging hashing, k-anonymity, and other established network data privacy solutions.

The main drawbacks of this approach are that i) outdoor usage can be a challenge unless cellular data is available or outdoor WiFi coverage is available, ii) accessing the WiFi data from new areas can be tedious if the required WiFi data expertise is not available (our team has a decade of experience working with both Cisco and Aruba solutions), iii) not every person in the area will have a WiFi-enabled device or have their WiFi interface turned on (note: the coverage will still be significantly better (at over 90% coverage for  and ) than any app-driven approach), and iv) WiFi MAC address anonymisation policies limit the information you can obtain from unassociated WiFi devices (the anonymisation prevents mobility analysis of the devices as their MAC addresses change, however occupancy analysis can still be performed).

Overall, we believe this is a very promising avenue to pursue and the administrators at all three universities agree. We are already providing updates at  to the campus facilities managers and the deans of students at  about the occupancy and mobility levels across the buildings and dorms at each campus.

5.4. Limitations

Our data-driven study had the following limitations. First, the data was only collected from WiFi-enabled devices that were associated with the campus WiFi networks. This can result in under-counting if individuals do not enable or carry a WiFi device or if they were not connected to the campus network. Second, our data comes from two campuses described, one each in the USA and Singapore. While we have observed similar results in two different settings in different continents, local factors may prevent specific results from applying to other regions. Still, we believe many of our observations about local and mobility spread resulting from quarantine and social distancing policies are more broadly applicable.

6. Related Work

This analysis was inspired by prior work that uses WiFi data to analyse and understand other environments. This includes understanding the flow of people through buildings (Trivedi et al., 2020), how people move in groups (Sen et al., 2014; Jayarajah et al., 2015), shopper behaviour (Hwang and Jang, 2017), and even the psychological state of individuals (Zakaria et al., 2019). We use similar techniques as prior work and apply it to understand the occupancy and mobility patterns of staff and students across our three campuses.

Our goal is to help advise policy makers about the effect of various COVID-19 related decisions. This is particularly important as every country is grappling with COVID-19 and enacting a slew of policies that they believe are best for them (Organization, 2020a). But how do you know if these policies are effective? To help, there have been several online COVID-19 trackers that show the number of infections per country (Worldometer, 2020; Organization, 2020b;, 2020; for Systems Science and at Johns Hopkins University (JHU), 2020). In addition, there has also been more focused COVID-19 analysis such as the analysis of an office block in South Korea (Park et al., 2020), analysis of possible ways to exit lockdown across the world (YAP et al., 2020), the rate of spread of COVID-19 in China (Surveillances, 2020), and brief analysis of campus behaviour in Italy (Favale et al., 2020). We add to this growing body of work by providing detailed analysis using data from three different campuses (two of which are quite large) that shows the clear effect that various policies have, over two months, on the occupancy and mobility at these campuses.

7. Conclusions

In this paper, we present results from two campuses in Singapore and one in the North-Eastern portion of the United States of America showing the effect of various COVID-19 related containment policies on the occupancy and mobility patterns on each campus. Our data was collected directly from the WiFi infrastructure at each campus. Our results showed that work-from-home and split-team work arrangements were quite effective at reducing the overall occupancy of buildings on campus. However, reducing the mobility rate of individuals on campus require nothing less than a full lockdown.

Our analytics are already being used by campus administrators to understand the impact of the various policies being enacted. We are currently improving our analytics, user interfaces, and visualisations to allow administrators to easily understand the occupancy and mobility impact of various COVID-19 related policies such as mandatory social distancing, online learning, flipped classrooms, and split-team arrangements. This is particularly important as these measures will be used at scale when lockdowns are lifted, and classes resume for the Fall semester.