Tool Developed at Columbia Tracks COVID-19 as New York’s Lockdown Eases
A statistical programming language called Stan, created in 2012 by a team led by Columbia University’s Andrew Gelman, helps epidemiologists model how many lives are at risk of COVID-19 in New York—and when hospitals will be overrun with patients with the lockdown ending.
Stan performs Bayesian inference, which is a statistical method for combining information from multiple sources. Developers from around the world continually tweak the open source program to be ever more accurate. Now, New York Governor Andrew Cuomo’s advisors on COVID-19 data are using the program.
“Stan will help us approach the COVID crisis from a data-driven perspective and to reach a scientific consensus on the coronavirus,” said Samir Bhatt, a senior lecturer in geostatistics at Imperial College London whose team advises the Governor. “And that knowledge can than be used by Gov. Cuomo to make the most informed decisions.”
Bhatt gives Gelman and the Stan team credit for building a program that helps researchers make inferences under uncertainty as the lockdown lifts.