Itsik Pe'er - May 6, 2020

Itsik Pe'er, PhD, "Columbia University Effort to Predict Infections with Data (CUEPID)"

Professor Itsik Pe’er presented the results of a short term COVID-19 Data Challenge hosted by the Center for Health Analytics at Columbia's Data Science Institute. Final submissions were accepted through April 24th, and the winning teams were awarded cash prizes. The first track was to predict the cumulative number of confirmed cases, recovered cases and deaths for May 1, 2020. The 17 teams that participated leveraged diverse computational approaches, including linear regression, neural networks, and epidemiological modeling. The winning submission used a carefully crafted geometric time series model, and the second place submission was based on an SIR-type model. The second track was for an open-ended, creative analysis of COVID-19 data. The winning projects included a prediction of SARS-CoV-2 peptides that elicit strong T-cell response, as well as an analysis of social distancing, mobility and public policy. Projects that came in second place included an analysis of predictors of social distancing adherence, and an algorithmically rendered musical composition based on the Coronavirus RNA genome. More information about the Data Science Institute’s COVID-19 efforts can be found here: https://datascience.columbia.edu/dsi-...