Visual Analytics for Explaining and Analysing Contact Tracing Networks
Lead Research Organisation:
Swansea University
Department Name: College of Science
Abstract
Contact tracing networks carry invaluable information for researchers to understand the spread of the virus, for policy-makers to control the COVID-19 outbreak, and for the government and the media in informing the public in rich ways. However, current data science tools fall short for the exploratory and explanatory analysis of the temporal, spatial and social aspects of these networks, and little is known on how most effectively the results of such analyses can be communicated broadly. This lack of a toolbox leads to organisations wasting resources on developing partial solutions designed without broad stakeholder engagement. To this end, this project aims to follow a user-centred approach to develop visual analytics methods for the analysis of large collections of contact tracing networks along with techniques for the communication of analysis results in transparent, comprehensive, yet engaging ways. Contact networks come with noteworthy technical and ethical challenges: inherent uncertainties due to the variation in their generation mechanisms, e.g., apps, hospital records, by volunteers; and high volumes of complex and sensitive information represented as event-based interactions with spatio-temporal facets. This project responds to these challenges through two deliverables comprising visualisation methods working simultaneously at group and individual levels while communicating the general trends in the spread:
1. Visualisations aimed at experts for understanding collections of contact networks to inform public health policies and make in-depth investigations without compromising individuals' privacy.
2. Visualisations for communicating analysis results with the general public for information and evidencing policy recommendations with representations having a purely explanatory emphasis.
1. Visualisations aimed at experts for understanding collections of contact networks to inform public health policies and make in-depth investigations without compromising individuals' privacy.
2. Visualisations for communicating analysis results with the general public for information and evidencing policy recommendations with representations having a purely explanatory emphasis.
Publications
Baumgartl T
(2021)
In Search of Patient Zero: Visual Analytics of Pathogen Transmission Pathways in Hospitals
in IEEE Transactions on Visualization and Computer Graphics
Lee A
(2021)
The Effectiveness of Interactive Visualization Techniques for Time Navigation of Dynamic Graphs on Large Displays.
in IEEE transactions on visualization and computer graphics
Sondag M
(2022)
Visual Analytics of Contact Tracing Policy Simulations During an Emergency Response
in Computer Graphics Forum
Wallinger M
(2022)
Edge-Path Bundling: A Less Ambiguous Edge Bundling Approach.
in IEEE transactions on visualization and computer graphics
Description | In 2021-2022 we have developed a tool, in conjunction with epidemiologists, to visualise contact tracing simulations and experiment with policies applied to those simulations. At the beginning of the project, epidemiologists developed a network-based model. https://github.com/ScottishCovidResponse/scrc-vis-modelling/tree/master/ContactTracing This model simulates COVID-19 on top of a temporal network - a network where humans were undergoing contacts at specific times. The result of this simulation is a forest of dynamic trees. One of the primary issues of this approach is that the simulation is stochastic and it is very difficult to compare two different policies on the same network. We initially applied a series of static and dynamic graph drawing techniques and analytical methods to help epidemiologists gain an overview of the disease progression of a single policy but it was only possible to compare multiple policies statistically: https://github.com/ScottishCovidResponse/scrc-vis-modelling/tree/master/ContactTracing/GraphVisualisation We then developed a post-processing and visualisation technique that can compare policies in richer ways. Our approach simulates contact tracing without restrictions and a secondary algorithm applies the policy so that different policies are applied to the same simulation, they can be compared. We created a visualisation tool around this approach: (note that this not a public link yet) and evaluated it with epidemiologists. The resulting paper was accepted at EuroVis 2022 and won a best paper honourable mention at the conference. We have developed a tool to explain contact tracing simulations for the general public: https://contact-viz.cim.warwick.ac.uk/particle-people/ This tool demonstrates in an intuitive way how diseases spread through a population. Given this simulation, and other standard visualisation techniques, we are developing a human-centred study to determine the understandability of such simulations with the general public and, after seeing the visualisation, how likely they are to comply with policy. This study will be run on a crowdsourcing platform (Prolific) to determine relevant results for members of the general public in the United Kingdom. We have had a second paper accepted at ACM CHI 2023. This paper is an engagement with Technocamps and explored how visualisation interfaces can help explain disease modelling concepts to children. These activities are currently reaching about 100 children a month in Wales. |
Exploitation Route | We have a published, award-winning paper at EuroVis 2022 We have a paper accepted in minor revision at ACM CHI 2023 A third publication is in preparation. As explained above, we have had some progress on introducing visualisation in the area of epidemiology. We are also engaging with the general public. |
Sectors | Digital/Communication/Information Technologies (including Software),Education,Healthcare |
Description | We are in the process of integrating a variety of visualisation approaches into contact tracing simulations associated with the SCRC. https://www.gla.ac.uk/research/az/scrc/ At this point, the approaches have been used to visualise the results of policy simulations in order to validate parameters and compare some scenarios. Before the pandemic, we had not collaborated with epidemiologists before. The researchers that we have engaged with through RAMPVis and our grant at the beginning of the project did not much experience with visualisation and did not really use these techniques and methods in their workflows. Our paper which developed visualisation techniques to understand contact tracing simulations was well received with extensive positive feedback in our evaluation. We have engaged with Technocamps through workshops for school children. This will spread knowledge of the fundamentals of disease spread and computational modelling of diseases. This workshop has reached hundreds of children in Wales. We have a ACM CHI paper in minor revision that discusses these contributions. |
First Year Of Impact | 2022 |
Sector | Digital/Communication/Information Technologies (including Software),Education,Healthcare |
Impact Types | Societal |
Description | Welsh Government Sêr Cymru III |
Amount | £107,010 (GBP) |
Organisation | Government of Wales |
Sector | Public |
Country | United Kingdom |
Start | 10/2020 |
End | 03/2021 |
Description | Presentation at the Turing Institute |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | We continue to present on the topic of the pandemic at various seminars and workshops. Results from the work produced by this grant were presented at the annual visualisation interest group seminar at the Turing Institute https://www.turing.ac.uk/research/interest-groups/visualization This seminar is attended by members of academia and industry and provides a good venue to increasing public understanding of this work. |
Year(s) Of Engagement Activity | 2021 |
Description | Royal Society Meeting (RSS) |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Other audiences |
Results and Impact | Cagatay also presented the results of our work at the Royal Statistical Society's 2021 meeting (RSS2021) with a title "Visualising contact tracing models for policy making" under the "Visualisation and visual analytics" session: https://virtual.oxfordabstracts.com/#/event/public/1880/program The event was attended by practitioners and researchers both with the industry and government and raised interest and follow-up discussions in applying data visualisation more broadly. |
Year(s) Of Engagement Activity | 2021 |
Description | Technocamps Workshops |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Schools |
Results and Impact | When we began the public engagement objective of this grant in the summer, we decided to contact Technocamps in order to engage with children on the topic of computational modelling. Technocamps is a public outreach organisation based in Swansea University (https://www.technocamps.com/en/). They primarily encourage computers science education in secondary schools in Wales. Through our initial demonstrations of our visualisations, it was immediately clear that there was significant interest in developing seminars around pandemic modelling and visualisation for children. The goal of these seminars were to be informative, but also to encourage the next generation of scientists to be come interested in this area. Technocamps has worked with us to design a workshop around epidemics and epidemic modelling to inspire children. We use particle people as a base for these workshops. https://contact-viz.cim.warwick.ac.uk/particle-people/ We have already run workshops at the Swansea Science Festival https://www.swansea.ac.uk/research/civic-mission/swansea-science-festival/ The four sessions attracted 21 students. This dissemination has had a impact in helping with student education. Over the next few weeks, these workshops will be piloted in Wales. This pilot has already attracted six schools and 180 students who will take part in these workshops. These workshops have been running for over a year now and creating an important impact in Wales and continues to be delivered to around 100 young people every month. This work has been reported on the paper: From Asymptomatics to Zombies: Visualization-Based Education of Disease Modeling for Children published at ACM CHI 2023. |
Year(s) Of Engagement Activity | 2021 |