Sorts Bhopal’s seven ghats into risk tiers on CPCB Class B standards for the BMC planning officer, with every score linked to its MPPCB report page.
The problem
With Durga Visarjan just 9 days away (20 Oct 2026), the Bhopal Municipal Corporation (BMC) planning officer must deploy limited municipal resources—aerators, trash skimmers, cleaning crews, and testing teams—across Bhopal’s 7 immersion ghats.
Today, this decision is made by memory and guesswork because 3 years of MPPCB lab data (2023–2025) is buried inside 300+ pages of dense, unread PDFs. Meanwhile, the water is in biological crisis: 42 of 116 Dissolved Oxygen tests failed bathing limits (hitting 0.0 mg/L at Bairagarh 4 times; Shahpura suffocating at 2.5 mg/L before immersion even begins), 100 of 112 BOD tests breached standards, and key sites (Bairagarh & Malikhedi) had entire 2025 Durga test tables completely omitted from official reports.
How it works
Ghat Watch — an operational decision-support tool built for the single workflow of the BMC planning officer:
- Automated Risk Triage: Ranks all 7 Bhopal ghats into clear operational tiers (Insufficient Data, High, Medium, Low) based strictly on CPCB Class B bathing standards.
- 100% Page-Level Traceability: Every score and breach citation links directly to the exact MPPCB report and page number (e.g., MPPCB 2025, Page 28), giving officers
verifiable justification for the Collector and Commissioner.
- Tactical Action Directives: Replaces generic charts with clear field orders per site: Aerate (oxygen crash), Clean (organic sludge/turbidity), Test First (missing
data), Sample Metals (Lead/Chromium spikes), and NGT Lake Ban containment.
- "What-If" 2026 Simulator: Allows officers to input fresh pre-visarjan lab readings on the fly to see how real-time intervention updates the risk profile.
- 1-Click Bilingual Sitrep: Instantly generates a clean, official Hindi and English executive brief ready to forward on WhatsApp to field supervisors and leadership.
How Claude was used
- Data Extraction & Invariant Testing: Used Claude Code to audit the messy CSVs and PDFs, standardizing non-numeric flags (BDL, , *, blanks) and writing strict
automated tests (test_invariants.py) to ensure no missing data was ever mistakenly scored as "clean."
- Rapid Full-Stack Implementation: Claude Code scaffolded the entire architecture—connecting the mathematical scoring formula to the FastAPI backend and rendering the
responsive Next.js command center in hours.
- Official Bilingual Policy Drafting: Claude generated the administrative Hindi/English prompt templates, ensuring executive briefs sound like professional government
sitreps rather than machine-translated jargon.
- Knowledge & Context Graphing: Leveraged Claude to map regulatory dependencies (NGT lake orders, Ramsar wetland criteria, CPCB limits) directly into the code and
audit trail.
Team
LOGORITHMS