Scores Bhopal’s seven immersion sites 0–100 from 1,524 MPPCB tests of DO, BOD and pH, with sliders that re-weight the ranking.
The problem
Every festival season (Ganesh & Durga Puja), idol immersions severely contaminate Bhopal's water bodies. While the Madhya Pradesh Pollution Control Board (MPPCB) samples water quality across pre, during, and post-immersion phases, the data remains buried in complex, dense 90-120 page annual PDF reports. Historically, these readings were never synthesized against CPCB Class B (outdoor bathing) standards or compared year-over-year.
With Durga Visarjan arriving on 20 October 2026, Bhopal Municipal Corporation (BMC) planning officers face acute operational ambiguity: they lack an empirical, evidence-based prioritization system to decide which ghats need urgent aeration, containment booms, heavy trash skimmers, and post-immersion cleanup crews. Furthermore, following the NGT order banning immersions in the Upper and Lower Lakes, BMC urgently needed an operational compliance monitoring system for newly constructed artificial kunds.
How it works
We built 'Visarjan Ghat Watch' — a full-stack, 100% dynamic environmental decision-support system grounded in 1,524 certified MPPCB water quality tests across all 7 Bhopal immersion sites (2023–2025).
Key Capabilities:
- Multi-Criteria Risk Scoring (MCDA): Calculates an empirical 0-100 risk score based on DO deficit, BOD organic spike, and pH deviation, dynamically ranking ghats from Critical (#1 Bairagarh Kund) to Controlled (#7 Prempura).
- Live Sensitivity Sliders: Allows administrative officers to adjust parameter weights (DO, BOD, pH) in real-time, instantly re-ranking ghats without page reloads.
- 100% Auditability & Traceability: Every single reading links directly to its source MPPCB PDF report and exact page number.
- Dynamic Data & Site Onboarding: Includes a high-throughput CSV/Excel bulk ingestion engine supporting multi-parameter matrices, dynamic ghat registration with instant automated re-ranking, and a 1-click 'run.bat' launcher.
- Bilingual Grounded AI Chatbot: An intelligent assistant answering complex queries strictly verified against MPPCB official monitoring records.
How Claude was used
We utilized Claude as an autonomous full-stack pair programmer and data engineering architect throughout the hackathon lifecycle:
- Data Extraction & Cleansing: Prompted Claude to parse dense MPPCB annual PDFs, structure 1,524 historical water readings into tabular SQLite schemas, and cross-reference CPCB Class B thresholds.
- Mathematical Modeling: Collaborated with Claude to formulate the empirical MCDA risk weighting formula and design dynamic sensitivity sliders.
- Rapid Full-Stack UI/UX Engineering: Claude generated clean, accessible, bilingual (Hindi & English) responsive dashboards with seamless Chart.js and Leaflet GIS integrations.
- Edge-Case Hardening & Testing: Used Claude Code to build comprehensive automated test suites (test_system.py) verifying bulk multi-parameter CSV uploads and zero-downtime new ghat registrations.
Team
NeerRakshak