Ranks Bhopal’s ghats on one screen, separating water that was already bad before immersion from water that gets worse during it, with a reason for each.
The problem
MPPCB tests the water at Bhopal's immersion ghats every year, before, during and after immersion. But the results sit in reports of 90 to 120 pages, so nobody really uses them when planning. We took three years of these readings (2023 to 2025) and turned them into one screen. It shows which ghats need attention first and why.
How it works
It ranks the ghats and gives a short reason for each one in plain words. The main idea is that it separates two different problems:
- Water that was already bad before immersion started. That's a lake or kund problem and needs longer-term work.
- Water that gets worse during immersion. That needs action on the day itself, like aeration, cleaning or extra testing.
Every number links to the exact report and page it came from, and you can see the page image right in the app. Every ghat also has a confidence level, so a ghat with missing data doesn't look safe just because nobody tested it.
Key features
- A ranked list of ghats with tier (High/Medium/Low), the reason, and confidence
- Click any number to see the report page it came from
- A gaps panel showing which ghats and years have no data, including Arch Bridge, which is on the 2026 list but has never been tested
- A one-page brief in Hindi and English that can be copied or downloaded and forwarded
- Upload new readings and the ranking updates, with a table of what changed
- A test planner: "I can test 3 ghats before 20 Oct, which ones?"
- A repeat-or-one-off badge, e.g. "BOD above limit in 6 of 6 tables"
How Claude was used
I used Claude Code to help write the code, the tests and the first draft of the Hindi text. A team member then reviewed the Hindi. The app itself uses no AI when it runs. Every number comes straight from the scoring code, never from an AI model. It works with no internet and no API key.
Team
Secretos