Turns three years of buried water-test PDFs into a ranked list of Bhopal’s immersion ghats, with plain reasons, traceable numbers and a Hindi and English brief.
The problem
Every year, MPPCB tests the water at Bhopal’s idol-immersion ghats and buries the results in 100-page PDFs that nobody compares against legal limits or previous years, so the BMC officer planning Durga visarjan (20 October 2026) has to decide from memory where to send cleaning, aeration and extra testing.
How it works
We built a simple tool that turns three years of buried water-test PDFs into a ranked list of Bhopal's immersion ghats, with plain-words reasons, numbers traceable to the report page, and a Hindi and English brief, so the officer planning Durga visarjan can decide where to send cleaning and testing without relying on memory.
How Claude was used
We used Claude to understand the problem and the data traps before writing any code, such as treating blanks as missing rather than clean, handling BDL readings, and choosing CPCB class B as our baseline. We then used Claude Code CLI as our main building tool. It wrote and refactored the data-cleaning code, the scoring that ranks ghats and explains each rank, the FastAPI backend and the WhatsApp delivery of the brief, working directly in our repo from the terminal. Claude also helped draft the brief wording so it stays factual and makes no health claims. Our team reviewed every change, spot-checked numbers against the original PDFs, and had the Hindi text checked by a Hindi speaker.
Team
X-Factor