Open-Source Infrastructure for a Pandemic Response
The COVID-19 outbreak has produced an unprecedented wave of collaborative, open-source work. Researchers, public health authorities, journalists, and volunteer programmers are applying the tools and workflows of software development—version control, public repositories, and distributed contribution—to the problem of tracking and understanding the pandemic. The resulting projects range from curated global datasets to DIY medical device plans.
Global and Regional Case Tracking
One of the most widely cited open datasets comes from the Johns Hopkins University (JHU) Center for Systems Science and Engineering, which powers an interactive dashboard for real-time tracking of reported cases. As the team notes in The Lancet, the tool was built "to provide researchers, public health authorities, and the general public with a user-friendly tool to track the outbreak as it unfolds." Data is pulled primarily from DXY and cross-referenced with the WHO and other sources; it has often been faster than the WHO to report a country's first confirmed case.
The University of Washington's Institute for Health Metrics and Evaluation has released the nCoV2019 dataset, presented on a dashboard of its own. This dataset is more granular, containing per-patient records with dates of symptom onset and laboratory confirmation. It is designed to support epidemiological calculations such as the reproduction number and incubation period.
For the United States, the COVID-19 Tracker project is a leading source on testing and infection rates, born from concern over the CDC's lack of public testing data. Launched in early March by a partnership between The Atlantic and the founder of Related Sciences, the project rallied volunteers to build a suite of software that crawls state agency websites, aggregates the data, and exposes it via a public API, a Google Sheet, and a web page. Our World in Data now relies exclusively on these numbers for its COVID-19 testing page.
Several national and municipal governments have also open-sourced their official statistics. Italy publishes its daily updates on GitHub, feeding a national dashboard. Tokyo and Zurich similarly maintain public repositories of local infection data.
Scientific Modeling and Genomic Epidemiology
Beyond case counts, open collaboration is accelerating basic research. The distributed-computing project Folding@home has directed volunteer compute cycles toward modeling molecular dynamics for COVID-19 drug design, with the effort's data stored in a public repository. The project is fully open source.
Nextstrain takes an open-source approach to genomic epidemiology. Its dashboard visualizes the evolutionary relationships among HCoV-19 viral mutations, sourced from GISAID, to help trace the virus's origins. The project's stance is that "current scientific publishing practices hinder the rapid dissemination of epidemiologically relevant results," and it prioritizes speed of data release over traditional publication cycles.
Smaller research datasets are also proliferating, including a repository of chest X-ray images aimed at training diagnostic AI models.
Forecasting and Decision-Support Tools
A number of projects focus on modeling how the outbreak will strain health care systems. COVID-19 Scenarios is an outbreak simulator for assessing regional hospital capacity.
COVID-19 Dashboards provides interactive visualizations of JHU data, built using Jupyter Notebooks and fastpages. The site is updated automatically via GitHub Actions and, because it includes predictive models, benefits from a fully transparent pipeline where the source code is embedded directly in the published web pages.
For comparative analysis, Predict COVID-19 allows users to measure case trajectories across countries. Various lightweight APIs also simplify programmatic access to JHU case data and drive numerous third-party visualizations.
In the US, The New York Times recently published its own dataset compiled from state and county sources. COVID-19 India, built by volunteers, offers a dashboard hosted on GitHub Pages with its source code available for reuse. Ontario has likewise published the source for its self-assessment tool.
Volunteer Networks and Citizen Projects
The Wuhan2020 community project is a self-organized effort to build a real-time data service synchronizing information across hospitals, factories, and procurement channels. The World Healthcare Organization app collective, led by Dr. Daniel Kraft, is developing a mobile application designed to provide local, personalized guidance while feeding anonymized data back to public health officials.
On the hardware side, the Low-Cost Open-Source Ventilator project publishes detailed plans for building a low-cost respirator from readily available parts, a contingency measure for hospitals facing exhausted supplies.
The scale of contribution in this space is striking: for free and open-source projects alone, more than 6,000 contributors have supported over 3,000 COVID-19 response efforts, collectively accumulating over 2 million views from more than 140,000 users. The primary hub for this coordination is GitHub itself, where maintainers encourage tagging repos with the "covid-19" topic to improve discoverability.
