This series has been closed with a total of 14 webinars. It started on April 17, 2020 with Dr. Peter X.-K. Song and ended with on July 24 with Dr. Bin Yu. Over a thousands attendees registered cumulatively. We thank all the webinar presenters and attendees for joining us on this platform and contributing to the battle with COVID-19 through data science in action. Comments and suggestions are welcome to the organizers.
A special issue of the Journal of Data Science on "Data Science in Action in Response to the Outbreak of COVID-19" was released on July 21, 2020; see http://jds-online.org/. The issue features 8 articles, including two discussion papers. Authors of three papers presented part of their works in this series.
Friday, 07/24/2020, 11 am EDT, Dr. Bin Yu, University of California, Berkeley (Video | Slides):
Curating a COVID-19 data repository and forecasting county-level death counts in the United States
Friday, 07/17/2020, 11 am EDT, Dr. Nicholas Reich, University of Massachusetts, Amherst (Video | Slides):
The COVID-19 Forecast Hub: using statistics and data science to support decision-making in a pandemic
Friday, 07/10/2020, 11 am EDT, Dr. Yuan Ji, University of Chicago (Video | Slides):
Semiparametric Bayesian inference for the transmission dynamics of COVID-19 with a state-space model
Friday, 06/26/2020, 11 am EDT, Dr. Kimia Ghobadi, Johns Hopkins University (Video | Slides available from Dr. Ghobadi upon request):
The opportunities and challenges of healthcare systems in the COVID-19 era
Friday, 06/19/2020, 11 am EDT, Dr. Yifan Zhu, Fred Hutchinson Cancer Research Center (Video | Slides available from Dr. Zhu upon request):
A statistical transmission model for COVID-19 outbreak with adjustment of external factors
Friday, 06/12/2020, 11 am EDT, Dr. Dean Follmann, National Institute of Allergy and Infectious Diseases (NIAID) (Video | Slides):
Statistics and modeling in response to COVID-19 at NIAID
Friday, 06/05/2020, 11 am EDT, Dr. Usha Govindarajulu, Icahn School of Medicine at Mount Sinai (Video | Slides):
A biostatistician's encounter with COVID-19 in New York City
Friday, 05/29/2020, 11 am EDT, Dr. Grace Yi, University of Western Ontario (Video | Slides):
Can the reported COVID-19 data tell us the truth? Scrutinizing the data from the measurement error models perspective
Friday, 05/22/2020, 11 am EDT, COV-IND-19 Study Group ( (Dr. Bhramar Mukherjee, University of Michigan; Dr. Debashree Ray, Johns Hopkins University; Rupam Bhattacharyya, University of Michigan; and Maxwell Salvatore, University of Michigan) (Video | Slides):
Predictions, role of interventions and implications of a national lockdown on the COVID-19 outbreak in India
Friday, 05/15/2020, 11 am EDT, Dr. David Corliss, Peace-Work (Video | Slides):
COVID-19: Analytic studies and opportunities for outside of epidemiology
Friday, 05/08/2020, 11 am EDT, Dr. Jing Qin, National Institute of Allergy and Infectious Diseases (NIAID) (Video | Slides):
Estimation of incubation period distribution of COVID-19 using disease onset forward time: A novel cross-sectional and forward follow-up study
Friday, 05/01/2020, 11 am EDT, Dr. Lily Wang, Iowa State University (Video | Slides):
COVID19 US dashboard: Spatiotemporal dynamics, nowcasting and forecasting of COVID-19 in the United States
Friday, 04/24/2020, 10 am EDT, Dr. Song Xi Chen, Peking University (Video | Slides):
Tracking reproductivity of COVID-19 pandemic with varying coefficient epidemiological models
Friday, 04/17/2020, 11 am EDT, Dr. Peter X.K. Song University of Michigan (Video | Slides available from Dr. Song upon request):
An epidemiological forecast model to assess the effect of social distancing on flattening the coronavirus curve in the USA
We develop a health informatics toolbox that enables us to project time-course dynamics of the COVID-19 epidemics in the USA. This toolbox is built upon a hierarchical epidemiological forecast model for observed daily proportions of infected and removed cases that are generated from an underlying Markov process of evolving Susceptible, Infectious and Removed (SIR) compartments of the COVID-19 infectious disease. We extend the classical SIR model to incorporate various types of time-varying social distancing protocols, which allows us to assess the effect of social distancing on flattening the coronavirus curve in the US. Some possible extensions of the epidemiological model to predict county-level risk are discussed. Such regional risk information is of critical importance for business reopening in the near future.
The first component of this talk is in a forthcoming discussion paper in Journal of Data Science (under "Accepted Papers").
Peter Song is a Professor of Biostatistics at the Department of Biostatistics, School of Public Health, University of Michigan. He received his PhD in Statistics from the University of British Columbia in 1996. Prior to the appointment at the University of Michigan, he was a faculty member at the Department of Statistics and Actuarial Science, University of Waterloo (2004-2007) and a faculty member at the Department of Mathematics and Statistics, York University, Toronto (1996-2004). Peter Song's research interests include bioinformatics, longitudinal data analysis, missing data problems in clinical trials, statistical genetics, and time series analysis. He is interested in methodological developments related to modelling, statistical inference and applications in biomedical sciences. In particular, Dr. Song's research projects are strongly motivated from real world data analysis. In 2007 he published a monograph "Correlated Data Analysis: Modeling, Analytics and Applications" by Springer. Dr. Song is a Fellow of the American Statistical Association and an Elected Member of the International Statistical Institute.
We propose a varying coefficient Susceptible-Infected-Removal (vSIR) model that allows changing infection and removal rates for the latest corona virus (COVID-19) outbreak in China. The vSIR model together with proposed estimation procedures allow one to track the reproductivity of the COVID-19 through time and to assess the effectiveness of the control measures implemented since Jan 23 2020 when the city of Wuhan was cut off followed by an extremely high level of self-isolation in the population. Our study found that the reproductivity of COVID-19 has been significantly slowed down in the three weeks from January 27 to February 17th with 96.3% and 95.1% reductions in the effective reproduction numbers R among the 30 provinces and 15 Hubei cities, respectively. Predictions to the ending times and the total numbers of infected are made under three scenarios of the removal rates. Extensions to the more general SEIR model setting are discussed, which are used for modeling and assessment for international comparative studies.
A part of the presented work is a forthcoming discussion paper in Journal of Data Science.
Song-Xi Chen is a University Chair Professor at Guanghua School of Management and Center for Statistical Science, Peking University. Dr. Chen joined the Guanghua School of Management in May 2008 and is now Co-Chair of the Department of Business Statistics and Econometrics, and the founding Director of the Center for Statistical Science at Peking University. Dr. Chen graduated from Beijing Normal University with bachelor's degree in mathematics and a master's degree in Mathematical Statistics. In 1990 he received a master's degree in Statistics and Operations Research from Victoria University of Wellington. In 1993, he obtained a Ph.D in Statistics from Australian National University. Dr. Chen's research interests are high dimensional statistical inferece, air quality assessment, population census, and inference for stochastic processes. He teaches Advance Multivariate Analysis, Asymptotic Statistics and Applied Statistics in graduate programs.
In response to the ongoing public health emergency of COVID-19, we investigate the disease dynamics to understand the spread of COVID-19 in the United States. In particular, we focus on the spatiotemporal dynamics of the disease, accounting for the control measures, environmental effects, socioeconomic factors, health service resources, and demographic conditions that vary from different counties. In the modeling of an epidemic disease, mathematical models are useful to understand the evolution, simulate the effects of health interventions, and forecast the future course of the disease. However, pure mathematical modeling is deterministic, and only demonstrates the average behavior of the epidemic; thus, it is difficult to quantify the uncertainty. Instead, statistical models provide a rich characterization of different types of errors. In this project, we investigate the disease dynamics by working at the interface of theoretical models and empirical data by combining the advantages of mathematical and statistical models. We develop a novel nonparametric space-time disease transmission model for infection count data, to study the spatial-temporal pattern in the spread of COVID-19 at the county level. The proposed methodology can be used to dissect the spatial structure and dynamics of spread, as well as to assess how this outbreak may unfold through time and space. Based on our research findings, we established a dashboard (https://covid19.stat.iastate.edu/) with multiple R shiny apps embedded to provide a real-time forecast of COVID19 infection count and death count at both the county level and state level, as well as the corresponding risk analysis.
Lily Wang is a tenured Associate Professor of Statistics at Iowa State University. She received her Ph.D. in Statistics from Michigan State University in 2007. She was a tenure-track Assistant/tenured Associate Professor in the Department of Statistics at the University of Georgia 2007-2013/2013-2014. Dr. Wang joined the Department of Statistics at Iowa State University in 2014. Her primary areas of research include developing cutting-edge statistical non/semi-parametric methods, statistical (machine) learning, methodologies for functional data, imaging data, and spatiotemporal data, survey sampling, high dimensional data analysis, and the application of statistics to problems in economics, neuroimaging, transportation, genetics, official statistics, environmental studies, and biomedical science. Dr. Wang is an Elected Member of the International Statistical Institute.
The current outbreak of coronavirus disease 2019 (COVID-19) has quickly spread across coun- ties and become a global crisis. However, one of the most important clinical characteristics in epidemiology, the distribution of the incubation period, remains unclear. In the literature, there exist many studies on estimating COVID-19 incubation period, some reports even from the same research team may contradict each other. It is very challenging and expensive to collect the contact- tracing incubation period data. The most difficult problem is to identify the infection onset time, especially for whose with long incubation periods. The accuracy can be highly influenced by re- call bias and lack of judgement of exposure. On the other hand, the symptomatic onset times for those confirmed individuals can be easily ascertained. Based on the COVID-19 daily updates from provincial and municipal health commissions in China, we notice that there is an abundance of cases who asymptomatically left Wuhan, the epicenter of COVID-19, and developed symptoms outside Wuhan. Due to stringent lockdown policy imposed by Chinese government in early stage of this epidemic (starting from January 23, 2020), it is reasonable to assume that those cases were infected before their departure from Wuhan. The time difference between departure time and symptoms onset is the censored observations of their incubation period. In length bias and renewal theory, this is called forward time. By employing probability renewal theory, we can use the forward time to estimate the incubation period unbiasedly. Our model estimated that about 5% to 10% of COVID-19 patients have incubation periods no less than 14 days. If validated, this might have implications for the length of a quarantine period in regions with a severe epidemic. To take the fact that some people may contract COVID-19 disease when departing from Wuhan into account, we have conducted a sensitivity analysis. The estimated incubation period differs by one to two days.
Jing Qin is a Mathematical Statistician at Biostatistics Research Branch in National Institute of Allergy and Infectious Diseases. After graduating from University of Waterloo (1992), he spent one year as a postdoctoral fellow at Stanford University before joining the faculty at the University of Maryland. Before moving to the National Institute of Health (2004), he worked at the Memorial Sloan-Kettering Cancer Center for 5 years. Dr. Qin’s research interests include the empirical likelihood method, case-control study, length bias sampling, econometrics, survival analysis, missing data, causal inference, genetic mixture models, generalized linear models, survey sampling and microarray data analysis. He was elected as a Fellow of the American Statistical Association in 2006. He is the author of "Biased sampling, over-identified parameter problems and beyond” (2017 Springer).
Like all pandemics, COVID-19 has many dimensions. Beyond the medical crisis, research is being conducted in a wide variety of disciplines.This has created many opportunities for statisticians and data scientists to contribute in areas outside of medicine and public health. This presentation gives an overview of some of the different areas where statisticians and data scientists are getting involved in the COVID-19 response, including economic impacts, sociological effects, supporting local schools and government, fact-checking, and the way we live and work.
With a background in statistical astrophysics, David Corliss leads an analytic team in industry with more than 20 years experience, especially time series methods in marketing and econometrics. He is active in the American Statistical Association, serving as Vice Chair of the steering committee of the Conference on Statistical Practice and writing Stats4Good, a monthly column on Data For Good for Amstat News. Dr. Corliss is the Founder and Director of Peace-Work, a volunteer cooperative of statisticians and data scientists providing analytic support for charitable groups and applying statistical methods to issue-driven advocacy in Data For Good projects.
India has taken strong and early public health measures for arresting the spread of the COVID-19 epidemic. With only 536 COVID-19 cases and 11 fatalities, India –-- a democracy of 1.34 billion people --– took the historic decision of a 21-day national lockdown on March 25. The lockdown was further extended to May 3rd, soon after the analysis of this paper was completed. The lockdown was again extended to May 18 while this paper was being revised.
In this paper, we use a Bayesian extension of the Susceptible-Infected-Removed (eSIR) model designed for intervention forecasting to study the short- and long-term impact of an initial 21-day lockdown on the total number of COVID-19 infections in India compared to other less severe non-pharmaceutical interventions. We compare effects of hypothetical durations of the lockdown on reducing the number of active and new infections. We find that the lockdown, if implemented correctly, has a high chance of reducing the total number of COVID-19 infected cases in the short term, and buy India invaluable time to prepare its healthcare and disease monitoring system. Our analysis shows we need to have some measures of suppression in place after the lockdown for increased benefit (as measured in terms of reducing the number of active cases). From an epidemiological perspective, a longer lockdown between 42-56 days is preferable to substantially "flatten the curve" when compared to 21-28 days of lockdown. Our models focus solely on projecting the number of COVID-19 infections and thus, inform policymakers about one aspect of this multi-faceted decision-making problem. We recognize that the collateral damage of a lockdown from social and economic perspective could be massive.
We conclude with a discussion on the pivotal role of increased testing, reliable and transparent data, proper uncertainty quantification, accurate interpretation of forecasting models, reproducible data science methods and tools that can enable data-driven policymaking during a pandemic. Our contribution to data science includes an interactive and dynamic RShiny app (covind19.org) with short- and long-term projections updated daily that can help inform policy and practice related to COVID-19 in India. We make our prediction code freely available for reproducible science and for other researchers to use these tools for their own prediction and data visualization work.
Bhramar Mukherjee is John D. Kalbfleisch Collegiate Professor and Chair, Department of Biostatistics; Professor, Department of Epidemiology, University of Michigan (UM) School of Public Health; Research Professor and Core Faculty Member, Michigan Institute of Data Science (MIDAS), University of Michigan. She also serves as the Associate Director for Quantitative Data Sciences, The University of Michigan Rogel Cancer Center. She is the cohort development core /co-director in the University of Michigan’s institution-wide Precision Health Initiative. Her research interests include statistical methods for analysis of electronic health records, studies of gene-environment interaction, Bayesian methods, shrinkage estimation, analysis of multiple pollutants. Collaborative areas are mainly in cancer, cardiovascular diseases, reproductive health, exposure science and environmental epidemiology. She has co-authored more than 240 publications in statistics, biostatistics, medicine and public health and is serving as PI on NSF and NIH funded methodology grants. She is the founding director of the University of Michigan’s summer institute on Big Data. Dr. Muhkerjee is a fellow of the American Statistical Association and the American Association for the Advancement of Science. She is the recipient of many awards for her scholarship, service and teaching at the University of Michigan and beyond, including the Gertrude Cox Award, from the Washington Statistical Society in 2016 and most recently the 2020 L.Adrienne Cupples Award from Boston University School of Public Health.
Bebashree Ray is an Assistant Professor of Epidemiology and Biostatistics in the Johns Hopkins Bloomberg School of Public Health. Her research focuses on developing novel statistical methods and software for discovering genetic determinants of common human diseases. She has primarily worked on type 2 diabetes, cardiovascular traits and orofacial clefts. Her research interests also include statistical methods for meta-analysis of cohorts, case-control studies, and multivariate analysis.
Rupam Bhattacharyya is a second-year PhD student in the Department of Biostatistics at the University of Michigan. He received his Master of Statistics and Bachelor of Statistics degrees from the Indian Statistical Institute, Kolkata. Currently progressing towards his candidacy, Rupam works on development and application of Bayesian methods in integrative omics and precision oncology with Prof. Veerabhadran Baladandayuthapani at the University of Michigan. Though his primary research interests circle around precision medicine and cancer research, Rupam is also excited about applied research in other wings of Biostatistics, including statistical genetics, disease epidemiology and clinical trials.
Maxwell Salvatore is a research area specialist in the Department of Biostatistics at the University of Michigan. He received his MPH in Epidemiology from the University of Michigan in 2017. He was advised by Dr. Rafael Meza and worked on liver cancer incidence trend analyses, while taking coursework in global health, cancer epidemiology, and modeling. Since then, he has been working with Dr. Bhramar Mukherjee on projects related to biobank-based research using Michigan Genomics Initiative and UK Biobank data, cancer risk and prevention, and health disparities. He is about to start as a doctoral student in Fall of 2020 in the Department of Epidemiology, University of Michigan School of Public Health.
The mystery of the coronavirus disease 2019 (COVID-19) and the lack of effective treatment for COVID-19 have presented a strikingly negative impact on public health. While research on COVID-19 has been ramping up rapidly, a very important yet overlooked challenge is on the quality and unique features of COVID-19 data. The manifestations of COVID-19 are not yet well understood. The swift spread of the virus is largely attributed to its stealthy transmissions in which infected patients may be asymptomatic or exhibit only flu-like symptoms in the early stage. Due to the limited test resources and a good portion of asymptomatic infections, the confirmed cases are typically under-reported, error-contaminated, and involved with substantial noise. If the drastic effects of faulty data are not being addressed, analysis results of the COVID-19 data can be seriously biased.
In this talk, I will discuss the issues induced from faulty COVID-19 data and how they may challenge inferential procedures. I will describe a strategy of employing measurement error models to address the error effects. Sensitivity analyses will be conducted to quantify the impact of faulty data for different scenarios. In addition, I will present a website of COVID-19 Canada (https://covid-19-canada.uwo.ca/), developed by the team co-led by Dr. Wenqing He and myself, which provides comprehensive and real-time visualization of the Canadian COVID-19 data.
Grace Y. Yi is a professor of the Department Statistical and Actuarial Sciences and the Department of Computer Science at the University of Western Ontario where she currently holds a Tier I Canada Research Chair in Data Science. Dr. Yi's research interests focus on developing methodology to address challenges concerning measurement error, causal inference, imaging data, missing data, high dimensional data, survival data, and longitudinal data. She authored the manuscript “Statistical Analysis with Measurement Error or Misclassification: Strategy, Method and Application” (2017, Springer). Dr. Yi received her Ph.D. in Statistics from the University of Toronto in 2000 and then joined the University of Waterloo as a postdoctoral fellow (2000-2001), Assistant Professor (2001-2004), Associate Professor (2004-2010), Professor (2010-2019), and University Research Chair (2011-2018). She is a Fellow of the Institute of Mathematical Statistics, Fellow of the American Statistical Association, and an Elected Member of the International Statistical Institute. In 2010 Dr. Yi received the Centre de Recherches Mathmatiques and the Statistical Society of Canada (CRM-SSC) Prize which recognizes a statistical scientist's excellence and accomplishments in research during the first fifteen years after earning their doctorate. She is a recipient of the University Faculty Award (2004-2009) granted by the Natural Sciences and Engineering Research Council of Canada. Dr. Yi’s work with Xianming Tan and Runze Li won The Canadian Journal of Statistics Award in 2016. Dr. Yi has served the professions in various capacities. She was the Editor-in-Chief of The Canadian Journal of Statistics (2016-2018), the President of the Biostatistics Section of the Statistical Society of Canada in 2016, and the Founder of the first chapter (Canada Chapter, established in 2012) of International Chinese Statistical Association. She will take on the Presidency of the Statistical Society of Canada for the period of 2020-2022.
In this presentation, I will discuss how working at a major New York City hospital system in the epicenter of the COVID-19 pandemic has changed my life forever. Around late March, I was pulled into COVID research for the hospital like I have never seen before. Suddenly being redeployed to work with a team of people who I had never met, who were also redeployed was a new challenge as well as keeping up with the daily download of updated data and analysis requests from infectious disease specialists to anesthesiologists. Dealing with real-time data analysis suddenly became the new normal. Handling messy data and constantly changing focus have been issues throughout. The challenge became being able to make meaning from all of this observational data. We certainly had more than enough patients but we needed to make correct interpretations of the messy and potentially biased data with meaningful statistical methods. Meanwhile there was institutional pressure to get the results fast. The pressure is still there and it is real as people are racing to find a cure for this horrible virus. Please hear how I have tried to manage during this crisis with making meaningful results out of data driven, time sensitive analyses.
Usha Govindarajulu is a Senior Faculty in the Center for Biostatistics in the Department of Population Health Sciences of the Icahn School of Medicine at Mount Sinai. She graduated from Boston University with a PhD in Biostatistics and spent two years as a postdoctoral fellow at Harvard School of Public Health. She then worked for a year research faculty at Yale University before moving back to Boston and working at Brigham & Women’s and Harvard Medical School. After being there about 5 years, she moved to New York and took as a position as an Assistant Professor of Biostatistics at SUNY Downstate School of Public Health. She was there approximately 7 years before leaving to be in her current position. Her research interests are in survival analysis, frailty models, causal inference, genetic epidemiology, and machine learning. She is the 2020 Chair-Elect of the Section on Statistical Computing of the American Statistical Association.
The COVID-19 pandemic is an enormous public health challenge which requires an extraordinary response. At the National Institute of Allergy and Infectious Diseases, statisticians have been deeply involved in crafting and executing research to better understand, treat and prevent COVID-19. This talk will focus on the statistical aspects of different current projects including design of vaccine trials, issues in constructing running platform treatment trials, epidemiologic modeling, and the role of the immune response in modulating vaccine efficacy.
Dr. Follmann is Chief of the Biostatistics Research Branch at the National Institute of Allergy and Infectious Diseases (NIAID), a role he has held for the past 16 years. He has authored or co-authored more than 250 peer-reviewed research articles and received numerous awards, including the Department of Health and Human Services Secretary’s Award for Distinguished Service, the Best Paper in Biometrics 2009, and is an elected Fellow of the American Statistical Association in 2003. He serves on committees and advisory boards for the US Food and Drug Administration, the National Institutes of Health, and academic departments. Current research interests focus on statistical methods related to vaccinology.
The COVID-19 pandemic showed some resemblances to previous outbreaks caused by coronaviruses such as SARS and MERS. However, by combing features such as relatively high transmission rate, potential of asymptomatic/pre-symptomatic infection and high case fatality rate among the vulnerable, this virus has presented significant challenges to public health systems. In this talk, I will introduce a statistical model for inferencing the transmission dynamics of COVID-19 outbreak incorporating some special features of the virus, and the approaches to combine certain spatial, temporal, social and demographic factors into the model. Early analytical results applied to the observed epi-curves in Wuhan, China enabled us to understand the outbreak monitoring data in other regions. We will present some preliminary inferences for the ongoing COVID-19 outbreak data from several US states.
Dr. Yifan Zhu is a staff scientist from the Vaccine and Infectious Disease Division, Fred Hutchinson Cancer Research Center. He received PhD in Biostatistics from University of Florida in 2015 before joining the Hutch. Dr. Zhu has worked on various projects on design and inference of HIV prevention/vaccine trials, infectious disease transmission dynamic models, physical activity monitoring studies, as well as methodological research such as Bayesian model selection, functional data analysis, and compositional data analysis.
The COVID-19 pandemic has been a major global challenge since its emergence. With the fast growth in its range of spread, the available data around the pandemic and its impacts are also growing at a fast pace. As the pandemic threatens to overwhelm healthcare systems and as hospitals and businesses start to reopen, how can we leverage data (with all its richness and restrictions) to strategize best responses in healthcare systems. In this talk, we will highlight an ensemble of the use of data to better understand our social and healthcare responses to the COVID-19 pandemic. For instance, how hospital operations can benefit from data and mathematical optimization. What are the healthcare systems' response to personal protective equipment (PPE) shortage, and how does it affect hospitals' precaution procedures. How hospitals can reopen safely while under continued pressure (and with the forecasts of upcoming surges). How system-level pooling strategies can be employed to efficiently allocate resources (e.g., beds, nurses, PPE, ventilators) and minimize adverse effects.
Dr. Kimia Ghobadi is a John C. Malone Assistant Professor of Civil and Systems Engineering and a member of the Malone Center for Engineering in Healthcare and the Center for Systems Science and Engineering (CSSE), Johns Hopkins University (JHU). Prior to joining JHU, she was a postdoctoral fellow at MIT Sloan School of Management and obtained her PhD from the University of Toronto in Industrial Engineering. Her research interests are in developing inverse and forward optimization models, real-time algorithms, and analytics technics with application in healthcare systems including healthcare operations and medical decision-making.
The outbreak of Coronavirus Disease 2019 (COVID-19) is an ongoing pandemic affecting over 200 countries and regions. Inference about the transmission dynamics of COVID-19 can provide important insights into the speed of disease spread and the effects of mitigation policies. We develop a novel Bayesian approach to such inference based on a probabilistic compartmental model and data of daily confirmed COVID-19 cases. In particular, we consider a probabilistic extension of the classical susceptible-infectious-recovered model, which takes into account undocumented infections and allows the epidemiological parameters to vary over time. We estimate the disease transmission rate via a Gaussian process prior, which captures nonlinear changes over time without the need of specific parametric assumptions. We utilize a parallel-tempering Markov chain Monte Carlo algorithm to efficiently sample from the highly correlated posterior space. Predictions for future observations are done by sampling from their posterior predictive distributions. Performance of the proposed approach is assessed using simulated datasets. Finally, our approach is applied to COVID-19 data from four states of the United States: Washington, New York, California, and Illinois. An R package BaySIR is made available at https://github.com/tianjianzhou/BaySIR for the public to conduct independent analysis or reproduce the results in this paper at https://arxiv.org/abs/2006.05581.
Dr. Yuan Ji is a Professor of Biostatistics in the Department of Public Health Sciences, University of Chicago. He received his PhD in Statistics from University of Wisconsin-Madison in 2003. He was assistant and associate professor of Biostatistics at the MD Anderson Cancer before moving to Chicago in 2012, where he directed the Program of Computational Genomics & Medicine, Northshore University HealthSystem for 6 years. His research areas are innovative designs for clinical trials that aim to improve efficiency in practice, and novel statistical tools for high dimensional informatics data. Dr. Ji is a Fellow of the American Statistical Association.
The COVID-19 Forecast Hub is a consortium of modeling teams from across the world making forecasts of the COVID-19 pandemic in the US. Launched in March 2020, the Hub has aggregated over 40m rows of data from over 40 different models. The forecast data are available publicly, and are fed weekly to the CDC's COVID-19 forecasting website: https://www.cdc.gov/coronavirus/2019-ncov/covid-data/forecasting-us.html. In addition to serving as a central repository of forecast data, the Hub also actively develops and releases weekly an ensemble forecast that combines together a subset of the submitted models. This talk will describe the process of building and maintaining the Hub, from the data model used to represent forecasts to the statistical challenges in building and evaluating an ensemble forecast in real-time. More information about the COVID-19 Forecast Hub can be found at https://covid19forecasthub.org .
Dr. Nicholas G Reich is an Associate Professor of Biostatistics at the University of Massachusetts, Amherst. He received a PhD in Biostatistics from Johns Hopkins School of Public Health. His research team at UMass has developed statistical methods and open-source tools for creating probabilistic, ensemble forecasts of infectious disease outbreaks in real-time. His team leads two international infectious disease forecasting consortia, including the FluSight Network and the COVID-19 Forecast Hub. Dr. Reich is the director of an Influenza Forecasting Center of Excellence, funded by the U.S. Centers for Disease Control and Prevention (CDC). Read more about his research lab at https://reichlab.io.
As the COVID-19 outbreak continues to evolve, accurate forecasting continues to play an extremely important role in informing policy decisions. In this talk, I will describe a large data repository containing COVID-19 information curated from a range of different sources. This data is then used to develop several predictors and prediction intervals for forecasting the short-term (e.g., over the next week) trajectory of COVID-19-related recorded deaths at the county-level in the United States.
Dr. Bin Yu is Chancellor’s Distinguished Professor and Class of 1936 Second Chair in the Departments of Statistics and of Electrical Engineering & Computer Sciences at the University of California at Berkeley and a former chair of Statistics at UC Berkeley. Dr. Yu's research focuses on practice, algorithm, and theory of statistical machine learning and causal inference. Her group is engaged in interdisciplinary research with scientists from genomics, neuroscience, and precision medicine. In order to augment empirical evidence for decision-making, they are investigating methods/algorithms (and associated statistical inference problems) such as dictionary learning, non-negative matrix factorization (NMF), EM and deep learning (CNNs and LSTMs), and heterogeneous effect estimation in randomized experiments (X-learner). Their recent algorithms include staNMF for unsupervised learning, iterative Random Forests (iRF) and signed iRF (s-iRF) for discovering predictive and stable high-order interactions in supervised learning, contextual decomposition (CD) and aggregated contextual decomposition (ACD) for interpretation of Deep Neural Networks (DNNs). Dr. Yu is a member of the U.S. National Academy of Sciences and a fellow of the American Academy of Arts and Sciences. She was a Guggenheim Fellow in 2006, and the Tukey Memorial Lecturer of the Bernoulli Society in 2012. She was President of IMS (Institute of Mathematical Statistics) in 2013-2014 and the Rietz Lecturer of IMS in 2016. She received the E. L. Scott Award from COPSS (Committee of Presidents of Statistical Societies) in 2018.