I’ve been working on Foreclosure Data Hub, which aggregates foreclosure auction data across the US.
The interesting part has been the data pipeline rather than the website itself. Foreclosure information is scattered across county sites, auction platforms, and other sources, all with different formats and update schedules.
I’m currently pulling from 20+ sources, normalizing the records, deduplicating properties, and enriching them with property/location data. One surprisingly difficult problem has been handling stale records when upstream sources change or republish old auction information.
Still working on improving data freshness and coverage:
The interesting part has been the data pipeline rather than the website itself. Foreclosure information is scattered across county sites, auction platforms, and other sources, all with different formats and update schedules.
I’m currently pulling from 20+ sources, normalizing the records, deduplicating properties, and enriching them with property/location data. One surprisingly difficult problem has been handling stale records when upstream sources change or republish old auction information.
Still working on improving data freshness and coverage:
https://www.foreclosuredatahub.com/
Would be interested to hear from anyone who’s worked on aggregating messy public datasets at scale.