EcoSense is a public record of environmental contamination. It compiles peer-reviewed measurements and data from 20 monitoring programs — from PFAS in drinking water to trace metals in soil — and publishes every row with its source, its confidence and its coordinates.
A single PFAS concentration from a Michigan well, one heavy-metal reading in a Karachi dust sample, a pesticide fingerprint in a Seine tributary — each reported in a separate PDF, database, or national bulletin, in its own units and its own language.
EcoSense reads these documents the way a reference librarian would — one at a time, carefully, cross-checked by a council of language models — and it files the findings somewhere you can cite them. Every number links back to a paper and a page. Every coordinate links back to a method. Every method links back to a program.
Numbers refresh on build. Live state is fetched from the/stats/overviewendpoint.
Each feed is polled on its own schedule — daily, weekly, or annually — and normalised against a shared schema. Every row retains a reference to its origin program.
Every data point is freely licensed for academic and research use. The map is interactive; the CSV and GeoJSON feeds are machine-readable. Cite any row by its source paper and page number.