COVID-19 Data Science and Behavior Links

This page collects links to online resources for non-medical, non-expert data-scientists interested in understanding the COVID-19 pandemic and human behavioral responses to same. I’m new to Git pages and GitHub generally and welcome suggestions, edits, additions, etc.

Data repositories

Disease spread/status

• Johns Hopkins Github repo for data and dashboard visualization

• Worldometers interactive data visualizations by country

• New York Times disease/death count data by state

• This Wikipedia article contains a table with number of tests conducted per country and results.

Human behavior datasets

• Dataset of major containment efforts by country and date

• Attitudes about COVID-19, Pew research center interactive

• International COVID-19 survey web page, with link to OSF

• Large dataset indicating tweet ID and sentiment score for COVID-19 tweets
  • A beginners resource to hydrating and using Tweet datasets

• Unacast visualizations showing social distancing metrics including travel distance data.

• Google mobility reports showing plots of mobility data by country and subregion.

Models

• This Kaggle COVID-19 page has links to datasets, code for models, and contests.

• Here is code to implement global forecast using SIR models

• R code and libraries for running/assessing SIR models.

UW-Madison-based resources:

• The UW Data-science coalition for COVID-19 has links to many resources to several resources including state and national data and visualizations. Here are some I thought particularly useful:
  • COVIDIFY—Python library for getting latest global data
  • Wisconsin COVID-19 status visualized
  • Data and visualizations mapping mobility changes at the county level across US

• Jerry Zhu’s “ten-hundred” plot to visualize exponential growth (Youtube explanation and Matlab code):

• Laurent Lessard’s JuPyTeR notebook for downloading/visualizing Johns Hopkins data

• Scott Sievert’s blog visualizing various aspects of disease spread

Visualizations and interactive demos

• Washington Post: How social distancing flattens the curve. This is a compelling visualization of SIR models, and why social distancing is the best intervention we currently have for limiting mass death in US.

• Interactive simulation modeling at Melting Asphalt. This fantastic interactive tools for SIR modeling gives a sense for how different assumptions about the disease do/do not impact total spread and fatalities, ending with an exercise in which you can try to figure out how best to “flatten the curve.”

• Video overview of SIR models and additional complexities from 3 brown 1 blue. This 20-minute video is worth your while if you want to understand some of the complexities that arise under SIR models with more realistic assumptions.

• Financial times COVID page with several graphical visualizations.

• An interactive visualization of the exponential spread of COVID-19 This visualization focuses on the growth curve (especially the exponential trend) of confirmed cases by country, US States.

Studies of human behavior.

• Vox article: Partisan affiliation is the best predictor of behavior during pandemic.
  • Primary-source paper.