Pervasive Scalable Technologies & Cyber-environment

We will use innovative computing to advance epidemiological science.

Key objectives:

Develop exa-scale enabled analytic methods and simulations of spreading process on multi-scale multi-layer (MSML) networks with a billion or more nodes.
Use simulation-based methods for analyzing complex interventions to contain real-time epidemics.
Combine simulations with HPC-enabled machine learning methods to support modeling and decision support tasks.

Scalable high-performance simulations, analytics, and ML.

Scalable epidemiological workflows to support COVID-19 planning and response

In [30] we present scalable high-performance computing-enabled workflows for COVID-19 pandemic planning and response. The scalability of our methodology allows us to run fine-grained simulations daily, and to generate county-level forecasts and other counter-factual analysis for each of the 50 states (and Washington, DC) and 3140 counties across the US. Our workflows use a hybrid cloud/cluster system utilizing a combination of local and remote cluster computing facilities, and using over 20,000 CPU cores running for 6–9 hours every day to meet this objective. Our models are used to guide COVID-19 planning and response efforts by public health agencies and universities in Virginia as well as by the DoD and CDC. We began executing these pipelines on March 25, 2020, and have delivered weekly updates to these stakeholders since that time without interruption.

Parallel algorithms for epidemic interventions

We study the problem of designing vaccination strategies in network models of epidemic spread satisfying the budget constraints to minimize the spread of an outbreak. This problem is computationally expensive even in a non-adaptive intervention setting where vaccinations are determined ahead of time and are not dependent on the current state of the epidemic, and vaccines are immediately effective with 100% efficacy. In [34], we use ideas from influence maximization for this problem. We observe that at low transmission probabilities, the number of nodes not infected is a submodular function. While this is not true in general, this motivates the use of nodes which maximize influence as a vaccination strategy. However, constructing such a set is inherently sequential since it is based on a greedy algorithm. We present a new parallel algorithm based on greedy hill climbing, and present an efficient parallel implementation for distributed CPU-GPU heterogeneous platforms. We show strong scaling results on up to 128 nodes of the Summit supercomputer. Our parallel implementation is able to significantly reduce time to solution, from hours to minutes on large networks.

Parallel epidemic diffusion simulator

UMD and UVA have teamed together to develop a new parallel epidemic diffusion simulator that uses a hybrid of time-stepping and discrete-event simulations, and is implemented on top of an asynchronous message-driven task-based parallel runtime called Charm++. The simulator will be modular and allow plugging in different models for disease progression, disease transmission, and interactions between agents. We have been able to implement the core simulation framework so far with support for real populations. The next steps involve performance analysis and improvements such as load balancing and communication optimization, and adding new features such as the ability to simulate interventions (such as vaccinations, school closures, contact tracing + quarantining, etc.).

References

[1] T. Abate. Crowdsourcing site collects county-level policy data to inform decisions about easing social- distancing, April 2020. https://news.stanford.edu/2020/04/13/stanford-crowdsources-county- level-covid-19-policy-data/.

[2] V. Abeykoon, N. Perera, C. Widanage, S. Kamburugamuve, T. A. Kanewala, H. Maithree, P. Wickramasinghe, A. Uyar, and G. Fox. Data engineering for hpc with python. In 2020 IEEE/ACM 9th Workshop on Python for High-Performance and Scientific Computing (PyHPC), pages 13–21. IEEE, 2020.

[3] A. Adiga, J. Chen, M. Marathe, H. Mortveit, S. Venkatramanan, and A. Vullikanti. Data-driven modeling for different stages of pandemic response. Journal of the Indian Institute of Science, pages 1–15, 2020.

[4] A. Adiga, D. Dubhashi, B. Lewis, M. Marathe, S. Venkatramanan, and A. Vullikanti. Mathematical models for covid-19 pandemic: a comparative analysis. Journal of the Indian Institute of Science, pages 1–15, 2020.

[5] A. Adiga, L. Wang, B. Hurt, A. Peddireddy, P. Porebski, S. Venkatramanan, B. Lewis, and M. Marathe. All models are useful: Bayesian ensembling for robust high resolution covid-19 forecasting, 2021. Sub- mitted.

[6] E. Bradley, M. Marathe, M. Moses, W. D. Gropp, and D. Lopresti. Pandemic informatics: Preparation, robustness, and resilience. arXiv preprint arXiv:2012.09300, published as part of CCC Quadrennial pa- pers, 2020. https://cra.org/ccc/resources/ccc-led-whitepapers/#2020-quadrennial-papers.

[7] B. Brubach, D. Chakrabarti, J. P. Dickerson, A. Srinivasan, and L. Tsepenekas. Fairness, semi- supervised learning, and more: A general framework for clustering with stochastic pairwise constraints. arXiv preprint arXiv:2103.02013, 2021.

[8] H. L. Carscadden, C. J. Kuhlman, M. V. Marathe, S. S. Ravi, and D. J. Rosenkrantz. Blocking the propagation of two simultaneous contagions over networks. In R. M. Benito, C. Cherifi, H. Cherifi, E. Moro, L. M. Rocha, and M. Sales-Pardo, editors, Proc. 9th International Conference on Complex Networks and Applications (Complex Networks), pages 455–468, Chan, Switzerland, 2020. Springer.

[9] S. Chang, E. Pierson, P. W. Koh, J. Gerardin, B. Redbird, D. Grusky, and J. Leskovec. Mobility network models of covid-19 explain inequities and inform reopening. Nature, 589(7840):82–87, 2020.

[10] J. Chen, S. Levin, S. Eubank, H. Mortveit, S. Venkatramanan, A. Vullikanti, and M. Marathe. Net- worked epidemiology for covid-19. Siam news, 2020.

[11] J. Chen, A. Vullikanti, S. Hoops, H. Mortveit, B. Lewis, S. Venkatramanan, W. You, S. Eubank, M. Marathe, C. Barrett, and M. A. Medical costs of keeping the us economy open during covid-19. Scientific reports, Oct 2020.

[12] J. Chen, A. Vullikanti, J. Santos, S. Venkatramanan, S. Hoops, H. Mortveit, B. Lewis, W. You, S. Eubank, M. Marathe, C. Barrett, and A. Marathe. Epidemiological and economic impact of covid-19 in the us. medRxiv : the preprint server for health sciences, November 2020.

[13] D. Chistikov, G. Lisowski, M. Paterson, and P. Turrini. Convergence of opinion diffusion is pspace-complete. In Proc. AAAI, pages 7103–7110. AAAI Press, 2020.

[14] E. Y. Cramer, V. K. Lopez, J. Niemi, G. E. George, J. C. Cegan, I. D. Dettwiller, W. P. England, M. W. Farthing, R. H. Hunter, B. Lafferty, et al. Evaluation of individual and ensemble probabilistic forecasts of covid-19 mortality in the us. medRxiv, 2021.

[15] S. Eubank, I. Eckstrand, B. Lewis, S. Venkatramanan, M. Marathe, and C. Barrett. Commentary on ferguson, et al.,“impact of non-pharmaceutical interventions (npis) to reduce covid-19 mortality and healthcare demand”. Bulletin of mathematical biology, 82(4):1–7, 2020.

[16] M. C. Fitzpatrick and A. P. Galvani. Optimizing age-specific vaccination. Science, 371(6532):890–891, 2021.

[17] G. Fox. Deep learning based time evolution, 2020. http://dsc.soic.indiana.edu/publications/ Summary-DeepLearningBasedTimeEvolution.pdf.

[18] G. Fox. Deep learning for spatial time series, 2020. http://dsc.soic.indiana.edu/publications/ Deep%20Learning%20for%20Spatial%20Time%20Series.pdf.

[19] G. C. Fox, G. von Laszewski, F. Wang, and S. Pyne. Aicov: An integrative deep learning framework for covid-19 forecasting with population covariates. arXiv preprint arXiv:2010.03757, 2020.

[20] F. Haghpanah, G. Lin, S. Levin, and E. Y. Klein. Analysis of the potential efficacy and timing of covid-19 vaccine on morbidity and mortality. Available at SSRN 3745195, 2021.

[21] A. Haque, M. Thakur, M. Bielskas, and A. V. A Marathe. Persistence of anti-vaccine sentiment in social networks through strategic interactions. In Thirty Fifth AAAI Conference on Artificial Intelligence (AAAI-21), 2021.

[22] D. Harris, T. Pensyl, A. Srinivasan, and K. Trinh. Dependent randomized rounding for clustering and partition systems with knapsack constraints. In International Conference on Artificial Intelligence and Statistics, pages 2273–2283. PMLR, 2020.

[23] P. J. Hotez, R. E. Cooney, R. M. Benjamin, N. T. Brewer, A. M. Buttenheim, T. Callaghan, A. Caplan, R. M. Carpiano, C. Clinton, R. DiResta, et al. Announcing the lancet commission on vaccine refusal, acceptance, and demand in the usa. The Lancet, 2021.

[24] K. Huang, T. Fu, W. Gao, Y. Zhao, Y. Roohani, J. Leskovec, C. W. Coley, C. Xiao, J. Sun, and M. Zitnik. Therapeutics data commons: Machine learning datasets and tasks for therapeutics. Nature Communications. Also as arXiv preprint arXiv:2102.09548, 2021.

[25] H. Kamarthi, L. Kong, A. Rodriguez, C. Zhang, and B. A. Prakash. When in doubt: Explainable and principled uncertainty quantification for epidemic forecasting. Submitted, 2021.

[26] X. Kang, S. Ranganathan, L. Kang, J. Gohlke, and X. Deng. Bayesian auxiliary variable model for birth records data with qualitative and quantitative responses. arXiv preprint arXiv:2008.06525, 2020.

[27] R. Laxminarayan, S. Jameel, and S. Sarkar. India’s battle against covid-19: Progress and challenges. The American Journal of Tropical Medicine and Hygiene, 103(4):1343, 2020.

[28] R. Laxminarayan, B. Wahl, S. R. Dudala, K. Gopal, S. Neelima, K. J. Reddy, J. Radhakrishnan, J. A. Lewnard, et al. Epidemiology and transmission dynamics of covid-19 in two indian states. Science, 370(6517):691–697, 2020.

[29] G. Lin, A. T. Strauss, M. Pinz, D. A. Martinez, K. K. Tseng, E. Schueller, O. Gatalo, Y. Yang, S. A. Levin, E. Y. Klein, et al. Explaining the “bomb-like” dynamics of covid-19 with modeling and the implications for policy. medRxiv, 2020.

[30] D. Machi, P. Bhattacharya, S. Hoops, J. Chen, H. Mortveit, S. Venkatramanan, B. Lewis, M. Wilson, A. Fadikar, T. Maiden, and C. L. Barrett. Scalable epidemiological workflows to support covid-19 planning and response. In In Proceedings of the 35th IEEE International Parallel and Distributed Processing Symposium, 2021.

[31] M. Marathe, A. Vullikanti, D. Rosenkratz, S. Ravi, R. Stearns, and S. Levin. Computational challenges and opportunities for forecasting epidemic dynamics using network models, 2020.

[32] N. Meghanathan. Exploring the step function distribution of the threshold fraction of adopted neighbors vs. minimum fraction of nodes as initial adopters to assess the cascade blocking intra-cluster density of complex real-world networks. Springer Applied Network Science (Published as part of Special issue on Epidemics Dynamics & Control on Networks), 5(97):1–33, December 2020.

[33] Z. Mehrab, A. G. Ranga, D. Sarkar, S. Venkatramanan, Y. C. Baek, S. Swarup, and M. V. Marathe. High resolution proximity statistics as early warning for us universities reopening during covid-19, 2020. medRxiv preprint 2020.11.21.20236042.

[34] M. Minutoli, P. Sambaturu, M. Halappanavar, A. Tumeo, A. Kalyanaraman, and A. Vullikanti. Preempt: Scalable epidemic interventions using submodular optimization on multi-gpu systems. In Proc. SC, SC ’20. IEEE Press, 2020.

[35] S. M. Moghadas, M. C. Fitzpatrick, P. Sah, A. Pandey, A. Shoukat, B. H. Singer, and A. P. Galvani. The implications of silent transmission for the control of covid-19 outbreaks. Proceedings of the National Academy of Sciences, 117(30):17513–17515, 2020.

[36] S. M. Moghadas, T. N. Vilches, K. Zhang, C. R. Wells, A. Shoukat, B. H. Singer, L. A. Meyers, K. M. Neuzil, J. M. Langley, M. C. Fitzpatrick, et al. The impact of vaccination on covid-19 outbreaks in the united states. medRxiv, 2020.

[37] A. Moitra and E. Mossel. Hot topics in computing: An invitation to computational epidemiology, March 2021.

[38] D. Morris, F. Rossine, J. Plotkin, and S. Levin. Optimal, near-optimal, and robust epidemic control. communication physics. OSF Preprints, 2020.

[39] Theoretically speaking — computational and statistical tools to control a pandemic: A panel discussion, simons institute, May 2020. https://simons.berkeley.edu/events/covid19.

[40] A. S. Peddireddy, D. Xie, P. Patil, M. L. Wilson, D. Machi, S. Venkatramanan, B. Klahn, P. Porebski, P. Bhattacharya, S. Dumbre, and M. Marathe. From 5vs to 6cs: Operationalizing epidemic data management with covid-19 surveillance. In Proceedings of the IEEE International Conference on Big Data (BigData), 2020.

[41] B. Peng, J. Li, S. Akkas, F. Wang, T. Araki, O. Yoshiyuki, and J. Qiu. Rank position forecasting in car racing. arXiv preprint arXiv:2010.01707, 2020.

[42] N. Perera, V. Abeykoon, C. Widanage, S. Kamburugamuve, T. A. Kanewala, P. Wickramasinghe, A. Uyar, H. Maithree, D. Lenadora, and G. Fox. A fast, scalable, universal approach for distributed data reductions. arXiv preprint arXiv:2010.14596, 2020.

[43] A. Pilehvari, W. You, J. Chen, S. Venkatramanan, J. Krulick, and A. Marathe. Differential impact of social distancing on covid-19 spread in the us: by rurality and social vulnerability, 2021. Submitted.

[44] J. D. Priest, M. V. Marathe, S. S. Ravi, D. J. Rosenkrantz, and R. E. Stearns. Evolution of similar configurations in graph dynamical systems. In R. M. Benito, C. Cherifi, H. Cherifi, E. Moro, L. M. Rocha, and M. Sales-Pardo, editors, Proc. 9th International Conference on Complex Networks and Applications (Complex Networks), pages 544–555, Chan, Switzerland, 2020. Springer.

[45] J. Quetzalc´oatl Toledo-Mar´ın, G. Fox, J. P. Sluka, and J. A. Glazier. Deep learning approaches to surrogates for solving the diffusion equation for mechanistic real-world simulations. arXiv e-prints, pages arXiv–2102, 2021.

[46] E. L. Ray, N. Wattanachit, J. Niemi, A. H. Kanji, K. House, E. Y. Cramer, J. Bracher, A. Zheng, T. K. Yamana, X. Xiong, et al. Ensemble forecasts of coronavirus disease 2019 (covid-19) in the us. MedRXiv, 2020.

[47] A. Rodr´ıguez, B. Adhikari, A. D. Gonz´alez, C. Nicholson, A. Vullikanti, and B. A. Prakash. Mapping network states using connectivity queries. Proceedings of IEEE International Conference on Big Data 2020 (IEEE BigData 2020), 2020.

[48] A. Rodr´ıguez, B. Adhikari, A. D. Gonz´alez, C. Nicholson, A. Vullikanti, and B. A. Prakash. Mapping network states using connectivity queries. NeurIPS 2020 Artificial Intelligence and Humanitarian and Disaster Relief (AI + HADR) Workshop, 2020.

[49] A. Rodr´ıguez, B. Adhikari, N. Ramakrishnan, and B. A. Prakash. Incorporating expert guidance in epidemic forecasting. arXiv preprint arXiv:2101.10247, 2020.

[50] A. Rodriguez, N. Muralidhar, B. Adhikari, A. Tabassum, N. Ramakrishnan, and B. A. Prakash. Steering a historical disease forecasting model under a pandemic: Case of flu and covid-19. NeurIPS 2020 Machine Learning in Public Health (MLPH) Workshop, 2020.

[51] A. Rodriguez, N. Muralidhar, B. Adhikari, A. Tabassum, N. Ramakrishnan, and B. A. Prakash. Steering a historical disease forecasting model under a pandemic: Case of flu and covid-19. Proceedings of the AAAI Conference on Artificial Intelligence, 2021.

[52] A. Rodriguez, A. Tabassum, J. Cui, J. Xie, J. Ho, P. Agarwal, B. Adhikari, and B. A. Prakash. Deep- COVID: An operational deep learning-driven framework for explainable real-time COVID-19 forecasting. In Proc. AAAI Conference on Artificial Intelligence, 2021.

[53] D. J. Rosenkrantz, M. V. Marathe, S. S. Ravi, and R. E. Stearns. Synchronous dynamical systems on directed acyclic graphs: Complexity and algorithms. In Proc. AAAI-2021 (to appear), 2021. 9 pages.

[54] C. Ruiz, M. Zitnik, and J. Leskovec. Identification of disease treatment mechanisms through the mul- tiscale interactome. bioRxiv, 2020.

[55] C. M. Saad-Roy, N. Arinaminpathy, N. S. Wingreen, S. A. Levin, J. M. Akey, and B. T. Grenfell. Implications of localized charge for human influenza a h1n1 hemagglutinin evolution: Insights from deep mutational scans. PLoS computational biology, 16(6):e1007892, 2020.

[56] C. M. Saad-Roy, B. T. Grenfell, S. A. Levin, L. Pellis, H. B. Stage, P. van den Driessche, and N. S. Wingreen. Superinfection and the evolution of an initial asymptomatic stage. Royal Society open science, 8(1):202212, 2021.

[57] C. M. Saad-Roy, S. A. Levin, C. J. E. Metcalf, and B. T. Grenfell. Trajectory of individual immunity and vaccination required for sars-cov-2 community immunity: a conceptual investigation. Journal of the Royal Society Interface, 18(175):20200683, 2021.

[58] C. M. Saad-Roy, S. E. Morris, C. J. E. Metcalf, M. J. Mina, R. E. Baker, J. Farrar, E. C. Holmes, O. Pybus, A. L. Graham, S. A. Levin, et al. Epidemiological and evolutionary considerations of sars- cov-2 vaccine dosing regimes. medRxiv, 2021. Science.

[59] C. M. Saad-Roy, C. E. Wagner, R. E. Baker, S. E. Morris, J. Farrar, A. L. Graham, S. A. Levin, M. J. Mina, C. J. E. Metcalf, and B. T. Grenfell. Immune life history, vaccination, and the dynamics of sars-cov-2 over the next 5 years. Science, 370(6518):811–818, 2020.

[60] Y. Serkez. The magic number for reducing infections and keeping businesses open, December 2020.

[61] A. Uyar, G. Gunduz, S. Kamburugamuve, P. Wickramasinghe, C. Widanage, K. Govindarajan, N. Per- era, V. Abeykoon, S. Akkas, and G. Fox. Twister2 cross-platform resource scheduler for big data, 2021.

[62] J. Vekemans, M. Hasso-Agopsowicz, G. Kang, W. P. Hausdorff, A. Fiore, E. Tayler, E. J. Klemm, R. Laxminarayan, P. Srikantiah, M. Friede, et al. Leveraging vaccines to reduce antibiotic use and prevent antimicrobial resistance: a who action framework. Clinical Infectious Diseases, 2021.

[63] C. E. Wagner, J. A. Prentice, C. M. Saad-Roy, L. Yang, B. T. Grenfell, S. A. Levin, and R. Laxmi- narayan. Economic and behavioral influencers of vaccination and antimicrobial use. Frontiers in public health, 8:975, 2020.
[64] F. Wang, C. Widanage, W. Liu, J. Li, X. Wang, H. Tang, and J. Fox. Privacy-preserving genomic computing with sgx-based big-data analytics framework. 20th IEEE International Workshop on High Performance Computational Biology (HiCOMB) [under review], 2021.

[65] L. Wang, A. Adiga, J. Chen, A. Sadilek, S. Venkatramanan, and M. Marathe. CausalGNN: Causal-based graph neural networks for spatio-temporal epidemic forecasting, 2021. Submitted.

[66] L. Wang, A. Adiga, S. Venkatramanan, J. Chen, B. Lewis, and M. Marathe. Examining deep learning models with multiple data sources for covid-19 forecasting, 2020. arXiv preprint arXiv:2010.14491.

[67] L. Wang, X. Ben, A. Adiga, A. Sadilek, A. Tendulkar, S. Venkatramanan, A. Vullikanti, G. Aggarwal, A. Talekar, J. Chen, B. Lewis, S. Swarup, A. Kapoor, M. Tambe, and M. Marathe. Using mobility data to understand and forecast covid-19 dynamics. In Proceedings of the 29th International Joint Conference on Artificial Intelligence Workshop on AI for Social Good., 2021.

[68] L. Wang, D. Ghosh, M. T. G. Diaz, A. K. Farahat, M. Alam, C. Gupta, J. Chen, and M. Marathe. Wisdom of the ensemble: Improving consistency of deep learning models. In NeurIPS, 2020.

[69] S. Weekes, P. March, M. Berger, M. Marathe, J. Crawley, C. M. Lois, A. El Bakry, R. Renaut, A. Gelb, F. Santosa, T. Grandine, P. Seshaiyer, and J. Hestheven. Report on future research directions for the national science foundation in the era of covid-19. A Report by SIAM Task Force on COVID-19, 2020. https://www.siam.org/Portals/0/reports/Report%20on%20Future% 20Research%20Directions%20for%20NSF.pdf?ver=2020-12-10-144744-750.

[70] Y. Wei and X. Deng. A simple gaussian process modeling approach for experiments with quantitative- sequence factors. Submitted, 2021.

[71] C. Widanage, N. Perera, V. Abeykoon, S. Kamburugamuve, T. A. Kanewala, H. Maithree, P. Wickramasinghe, A. Uyar, G. Gunduz, and G. Fox. High performance data engineering everywhere. In 2020 IEEE International Conference on Smart Data Services (SMDS), pages 122–132. IEEE, 2020.

[72] Q. Xiao, Y. Wang, A. Mandal, and X. Deng. Modeling and active learning for experiments with quantitative-sequence factors, revision for journal of the american statistical association. Submitted, 2021.

[73] O. Yagan, A. Sridhar, R. Eletreby, S. A. Levin, J. B. Plotkin, and H. V. Poor. Modeling and analysis of the spread of covid-19 under a multiple-strain model with mutations. Harvard Data Science Review, 2021.

[74] A. Yalaman, G. Basbug, C. Elgin, and A. P. Galvani. Cross-country evidence on the association between contact tracing and covid-19 case fatality rates. Scientific reports, 11(1):1–6, 2021.

A cyber-environment for real-time epidemic science: Team members have developed multiple dashboards to provide timely and accurate information to public health agencies, the general population, and the research community.

COVID-19 Surveillance Dashboard

At the beginning of the pandemic, there was a lot of confusion about where cases of COVID-19 were present and how prevalent the virus was. We initially developed the COVID-19 Surveillance Dashboard as a one-stop source where researchers and the public could easily visualize where cases were emerging and how many cases there were by region. Dashboard coverage rapidly grew to include all countries/territories that had confirmed cases, with province- or state-level detail for 20 countries, and county-level details for the US. This tool was the first dashboard we developed under this project and is the most popular of our dashboards: by October 2020, it had been visited by 1.13 million users from 220 countries.

The dashboard, which was released in early February 2020, initially covered only case counts as collated by Johns Hopkins University. Over time, however, we expanded to include additional open data sources, including the World Health Organization (WHO), Tencent, Wikipedia, and others; as new sources were added, some were retired, so our current data source dependencies include The New York Times, USAFacts, the COVID Tracking Project, the World Health Organization (WHO), Wikipedia, JHU (for Mainland China data), and Our World in Data for vaccination data. The dashboard is updated multiple times per day in order to provide users with the most up-to-date information.

As the pandemic progressed, additional features were added to the dashboard, including:

The ability to view counts of Confirmed Cases, Deaths, Estimated Active, and Estimated Recovered per day on the Heat Map, Chart, and Data views; users can also drill down to view these values for specific regions or as counts per 100K people in the population.
A timeline tool that allows users to view case counts for a specific date (default is current) and to cycle through the days of the pandemic to see how conditions have evolved over time. • Testing performed per day at the state level for the US; in addition to the total number of tests administered, the user can also see the number of positive and negative test results (note: not all values are available for all states; for example, Virginia only returns the number of positive test results).
The total number of vaccines administered by country (and, for the US, by state); this includes the number of vaccines administered, people who have received at least one vaccine, and people who are fully vaccinated.
An analytics tab where users can perform analytical queries, such as Top 6 deaths in world as a different way to focus in on the data that is of interest to them.

The front-facing COVID-19 Surveillance Dashboard is a Single Page Application which loads the current data into an HTML page for display and rendering, and the display is updated dynamically as the user interacts with the application. There are three main development APIs that are incorporated into the development of the dashboard: (i) ArcGIS API for performing the map rendering and functionality; (ii) Responsive Web Design to ensure usability across different displays, including mobile devices; and (iii) amCharts for the simple, yet flexible, charting capabilities. As the first of our dashboards to be developed, the COVID-19 Surveillance Dashboard framework was adopted as a standard in the development of our other dashboards, covered in the following sections.

Our paper describing our dashboard was presented at the IEEE International Conference on Big Data [40]. In addition to the history and technical details summarized above, the paper addresses some of the difficulties of using open data sources and proposes a standard for publishing open source data. It also describes in detail our algorithm for estimating recovery and active case counts, since this is important information for researchers that is often hard to obtain.

Medical Resource Demand Dashboard (MRDD)

One of the side effects of the COVID-19 pandemic is scarcity of medical resources: as more people become ill, hospitals can expect to see more patients, possibly exceeding their limited capacity. However, if public health officials can predict in advance where crises are likely to occur (and where they are less likely), they can prioritize distribution of ventilators and other medical supplies and personnel, strategize where field hospitals may be most effective, and essentially address the issue in a proactive, rather than reactive, way.

Our epidemiologists were already simulating projections of the number of hospitalizations that could be expected per region. Using hospital (bed) capacities by the Virginia Healthcare Alerting and Status System (VHASS) region for Virginia and by the Hospital Referral Region (HRR) for the US, we were able to use our hospitalization projections to estimate where hospitals could expect to exceed capacities, and provide a dashboard for public health officials where they could rapidly identify those regions on their own. Our dashboards (Virginia: https://nssac.bii.virginia.edu/covid-19/vmrddash and US: https://nssac.bii.virginia.edu/covid-19/usmrddash) provide the following visualizations:

The number of projected hospitalizations per region on a color-coded heat map.
The maximum projected percentage of occupied hospital beds per region on a color-coded heat map; we did not attempt to ascertain the occupancy per hospital because it was assumed that patients could be transferred between hospitals within a region relatively easily.
A timeline allowing users to update the heat maps from week-to-week over a six week period, with a movie feature that iterates through the weeks to allow users to see how hospital occupancies are expected to change over time.
A chart on the right side of the screen to allow users to see the time series of projected hospitalizations and hospital occupancy thresholds over a six-week period.

In order to translate the projected weekly hospitalizations into projected maximum weekly occupancy, we had to take a couple of additional factors into account:

It is unreasonable to assume that a hospital region’s entire capacity could be dedicated to COVID-19 patients; room must be spared for patients with other emergencies. Our initial assumption was that 80% of the hospital capacity would be taken up with non-COVID-19 patients, and that most hospital regions could accommodate up to 120% capacity with manageable effort. However, after the dashboards were originally deployed, we added a slider to allow users to change those settings in order to see how reducing the capacity reserved for non-COVID-19 patients (e.g., by cancelling elective surgeries) or adding additional beds (e.g., adding a field hospital) would impact occupancy.
The duration of people’s stays in the hospital also has a significant impact on occupancy levels. If a hospital that has the capacity for 50 COVID-19 patients gets 10 new patients per day, and the average stay is 2 days, then the hospital will never exceed capacity; however, if the average stay is 7 days, then occupancy will match or exceed capacity by the fifth day. In Virginia, the average duration was 8 days, while nationally the average stay was 7 days; we used these values as constants when we first released the dashboards, but subsequently released a feature which allowed users to adjust the average hospital stay in order to assess hospital occupancy using local hospital duration averages.

COVID-19 Mobility Impact Dashboard

The Stanford group combined a full mobility network, constructed from SafeGraph mobility data and demo- graphic data from American Community Survey, with an SEIR model to predict the spread of COVID-19 in 10 Metropolitan Statistical Areas (MSAs). What they found was that their mobility-network-enhanced SEIR model performed better in predicting spread of COVID-19 in those MSAs than either an aggregate-mobility- data-enhanced SEIR model or the SEIR model alone; it was responsive to traffic fluctuations to Points of Interest like restaurants, gyms, and churches; it predicted that minorities would suffer greater impact from the pandemic; and, most relevantly for our application, it found that targeted (strategic) closure of Points of Interest (POIs) could be more effective than uniform closures in slowing the disease spread. Reducing COVID-19 cases is our primary focus to be sure, but targeted closure also reduces the economic impact of the pandemic [9].

The MSAs that the Stanford group examined in their paper in Nature were large metropolitan areas, but the Virginia Department of Health (VDH) was interested in seeing if something similar could work at the state level. The UVA team partnered with Stanford’s team to recalibrate their model for three MSAs in Virginia, and used their projections to develop a dashboard where VDH could experiment with different mobility levels to try to ascertain where it would be most effective to reduce mobility. Stanford began by calibrating their model for three MSAs in Virginia:

Washington-Arlington-Alexandria-DC-VA-MD-WV (referred to here for brevity as Washington DC);
Richmond-VA (Richmond); and
Virginia-Beach-Norfolk-Newport-News-VA-NC (Eastern).

Simulations from their calibrated models for November 2020 to early January 2021 fit the actual COVID-19 activity over those two months very well.

For the dashboard, the Stanford team ran simulations for each of the three MSAs at foot traffic levels of 0%, 50%, 100%, and current levels relative to the corresponding foot traffic before the pandemic (e.g., comparing November 2020 foot traffic to November 2019 foot traffic) for each of five Point of Interest (POI) categories:

Restaurants, including full- and limited-service restaurants, snack bars, and cafes;
Essential Retail, including grocery stores, pharmacies, and convenience stores;
Retail, including clothing, hardware, book, and pet stores, and all other non-essential retail stores;
Gyms and Fitness centers; and
Religious Organizations, like churches, synagogues, and mosques.

The Dashboard allows users to visualize the impact of increasing or reducing foot traffic to different POIs on COVID-19 case counts. To this end, the dashboard is divided into 4 panels and has an additional popup report. The panel titled ”Visits to Points of Interest” on the left-hand side of the dashboard is the navigation bar of the application; users can see what the current foot traffic levels are for the selected region, and drag indicators along gliders to change the foot traffic mobility in POIs of that category. As they change the mobility levels on the navigation bar, the other three visible panels are updated accordingly. The upper right panel is the map panel; from here, users will see areas that show a decrease in case counts due to the mobility change highlighted in blue, and areas that show an increase in red; users can click on dates at the top of the bar to see cumulative case counts for one, two, three, and four week periods; finally, users can hover over areas on the map to see more details about the case counts, or select an area to highlight on the other two panels. The plot on the lower middle panel allows the user to view the cumulative case counts over the four week period at current and target mobility levels, along with uncertainty bounds. The table in the lower right panel shows the case count and percentage difference that would occur by MSA if foot traffic levels are changed. Finally, the user can click on the ”Mobility History” button on the navigation bar to display the ”Mobility data analysis for Virginia” report, which includes a full history of mobility in the three MSAs since the beginning of the pandemic, with a number of visualizations where users can review and make their own assessments. We submitted a paper based on this collaboration to KDD2021.

Unlike the surveillance dashboard and the MRDD, this dashboard is not open to the public. The reason for this is that a user viewing this tool in isolation might come to conclusions that are not fully informed, leading them to boycott certain POIs. Public health officials, on the other hand, can use this as part of a corpus of tools, and come to a more comprehensive conclusion regarding closures. For this reason, we maintain a password-protected site only for our colleagues in public health.