Post #1505388
2026-04-21 04:10 UTC
There's a key technical problem over at @civicband. I have blogged about it, and would appreciate reads and shares.
It's the most #CivicTech and #OpenData problem I've ever worked on.
https://meta.civic.band/archive/the-extraction-problem/
Replies (1)
-
@PeteBleackley@wandering.shop 2026-04-21 07:18
@phildini @civicband The approach I used when I did this for a client 1. Use OpenCV to split the table into cells 2. Run OCR on each cell individually 3. Use Pandas to assemble the contents of the cells into a dataframe