COVID-19 forecasting models

Interpretations from COVID-19 forecasting models make headlines nearly every day lately. To interpret the results of a predictive analytics model you have to understand the assumptions behind the model and its methods. I’m compiling a catalog of the more prominent models, along with ensemble visualizations that combine the forecasts of all of the models.

University of Texas COVID-19 forecasting model


        
          May 22, 2020

University of Texas COVID-19 model builds on the IHME model by adding real-time daily social-distancing data in order to project deaths from the first wave.

Imperial College of London COVID-19 forecasting models


        
          May 24, 2020

Arguably the most influential computer forecasting model in history, the first model from Imperial College was the one that led the UK and US governments to begin mitigation measures.

IHME COVID-19 forecasting model


        
          May 24, 2020

One of the most widely influential COVID-19 forecasting models has also been one of the most misinterpreted.

South Florida may see a new outbreak soon


        
          May 21, 2020

A model aimed at local forecasts projects a rapid spread of coronavirus in Miami-Dade and Broward counties, based on mobility, population density, age, insurance status, smoking prevalence, and weather.

covid19-models, mobility-data, unacast

Don't focus on the center line


        
          Apr 29, 2020

Don’t treat the center line of the IHME forecast charts as a precise prediction, any more than you would assume that a hurricane will follow the center line on the NHC hurricane forecast cone images. The forecast is a wide range, not a specific number.

covid19-models, ihme-model