FastOpenDataWe started FastOpenData after the third weekend in a row spent re-downloading TIGER shapefiles, untangling ACS column names, and re-implementing the same address-to-tract join for the same address-based product feature, on three different teams, at three different companies.
Every team thought they were the first to do it. Every team wrote their own crooked little ETL pipeline. Every team's pipeline broke when the Census Bureau renamed a directory. The data is public — has been for decades. The problem isn't the data. The problem is the work between you and the data.
"Hundreds of fields for any U.S. address. One API call. Stop reinventing the wheel."
So we built the join once, properly: weekly Planet OSM extracts piped through a Snakemake DAG, ACS tables fetched directly from Census APIs, USAspending contracts joined on tract centroids, all the way down to lead-paint exposure percentiles. Every field gets a stable name, a type signature, and a source attribution.
We charge for the API because servers cost money. The data itself is, and will always remain, public domain. If you ever want to leave, every field has a clear citation back to its original public source.