Everything you need to get started making models and visualizations using COVID-19 data

April 28th
(Brian McGowan/Unsplash)

This post is sponsored by Brilliant, which offers courses like Statistics Fundamentals, Probability Fundamentals, and Computer Science Fundamentals as well as more advanced courses like Introduction to Neural Networks, and Machine Learning. Sign up for Diagram today and get 20% off an annual membership.

COVID-19 is producing a vast wealth of data. Many groups, from hobbyists and insurance providers to governments and universities, have begun producing data to cover the crisis, create models and share information.

This is an ideal time for developers and machine learning students to get experience working with data to create dashboards, maps, and other visualizations using COVID-19 infection data. This post will show some inspiring uses of data, link to reliable data sources and list tools you can use to create work of your own.

Inspiration: data visualizations and models

General dashboards: coronastats.co, Johns Hopkins, COVID-19 spread sketch, Histogram of cases by US state, New York Times maps and timelines

Borders and quarantines: International border closures (disclosure: I made this), Global lockdown map

Projections: IHME, When the peak may come, by state

Other: Small business rescue success rates, by state

Data and APIs

COVID-19

The COVID Tracking Project: National and state data for the US. maintained by The Atlantic. Data contains:

  • Test results: positives, negatives and pending.
  • Current hospitalizations and cumulative.
  • Current ICU occupancy and cumulative.
  • Current ventilator use and cumulative.
  • Recovered and deaths.

World Health Organization: Global and regional data. Updated daily. Data contains:

  • New and cumulative positive cases.
  • New and cumulative deaths.
  • Transmission classifications (community transmission, clusters, etc.).
  • Days since last positive case, by country.

Johns Hopkins: Global and state, province, oblast, etc. data. National, state and county data in the United States. Data contains:

  • Positive and recovered cases by day.
  • Deaths by day.

New York Times: Global data. National, sate and county data in the United States. Data contains:

  • Positive cases and deaths.

Covid-19 Drive Thru Location Tracking: US data. Updated often. Data contains:

  • Location, hours of operations and any details about drive thru testing locations.

Borders, Economy and Supply chains

Marine Traffic: Live and archival global data. Data older than 24 hours is paywalled. Updated in real time. Data contains:

  • Ship locations, port of origin and destination.

FlightRadar24: Global flight data. Updated in real time. Data contains:

  • Plane type, airline and location, and departure airport and arrival airport.
  • Airline fleet and routes.
  • Aggregate flights made daily, with a rolling 7-day average.

IATA: International flight availability and border closure statuses, by country. Updated often. Data contains:

  • International flight options, by country.
  • Border open/closed status, by country

US Bureau of Economic Analysis: Global trade data detailing GDP and industries involved, by quarter and year. National and state data in the US. Updated quarterly. Data includes:

  • GDP, by country.
  • Personal income and employment in the US.
  • Import and export data between the United States and many countries.
  • US direct investment by industry, by country.

Tools

Map libraries and frameworks

Mapbox: an easy to use, flexible map maker you can embed in your site.

D3: a longtime JavaScript standard for building maps and other interactive data visualizations.

Tableau: a free and easy to use platform to visualize data.

Shareable, interactive dev environments

Glitch: develop sites, web apps, and visualizations under Glitch’s free hosting service.

Observable: turn JavaScript files into interactive pages. Break out functions into individual visualizations and interactive.

Machine learning platforms

TensorFlow: Google’s machine learning platform for JavaScript, Python, and other languages.

PyTorch: a TensorFlow alternative that’s gaining prominence. Developed by Facebook.

Databases and automation tools

Airtable: a spreadsheet tool with rich add-ons.

Zapier, Standard Library: automate your API calls in a visual tool, and connect digital services, like a spreadsheet to Webflow, easily.

NoCode tools

Webflow: a what you see is what you get design tool that creates surprisingly good websites without code.

Glide: create an Android or iOS app that pulls data from Google Spreadsheets.