// the take
A FastAPI-based REST API that aggregated COVID-19 data from JHU, CSBS, and NYT during the pandemic. It wrapped CSV data sources into a clean JSON API with per-location timelines, country filtering, and Swagger docs. Historically useful as a FastAPI reference project — the actual data it serves has been dead since 2023.
FastAPI usage is clean and idiomatic — the router separation between v1/v2 is a good pattern to copy. Multi-source abstraction (jhu/csbs/nyt each get their own service module) is well-structured for the problem. Test coverage is solid: separate test files per data source, fixture-based example data, and snapshot testing against expected JSON output. Docker support with both gunicorn and single-instance uvicorn Dockerfiles shows real deployment thinking.
The data is frozen — JHU stopped updating in 2023 and the live Heroku endpoint is almost certainly dead, so this is a read-only artifact now. No authentication or rate limiting on any endpoint; fine for a public COVID tracker but not a pattern to copy for anything that matters. The `recovered` field silently returns 0 because the upstream source dropped it — that kind of silent data failure with no indication in the API response is a real problem. Pipenv in 2024 is friction; the project would benefit from moving to pyproject.toml + uv, but nobody's maintaining it.
We read the README, the directory tree and a key source file in June 2026. Nobody pays to be listed. How we judge →