Progress: Researched a drinking water dashboard concept focused on mapping total dissolved solid (TDS) minerals like calcium, specifically for coffee and tea enthusiasts.
Struggle: Finding machine-readable drinking water data is kind of a pain. It seems to be locked in municipal- or watershed-based PDFs. I'm thinking this will require an OCR table extraction pipeline. So, I need to evaluate how much this could cost and the feasibility of using existing inference APIs or if it would be possible to run it locally with LocalLM or something similar.
Plan: More research on drinking water data availability, particularly in smaller municipalities (and is this different per state? (My guess is yes). I also have another idea for exploring whether raw water data (streams, wells, etc.) can be used for a different angle.