Method and limitations
How the ranking is built, what each state means, and what the data cannot tell you.
Known discrepancy in PRD v2 — needs a source-document correction
PRD v2 §6.5 states “High risk (6 of 7): all except Hathaikheda”. Its own rule applied to its per-ghat evidence gives 5 of 7: Khatlapura's lowest DO is 4.0 and 8 of 18 readings were below 5, so it meets neither condition. The backend applies the stated rule and also gives 5 of 7. The PRD has not been edited here; its owner should correct §6.5. This is not a new classification rule.
Criteria
CPCB Class B (outdoor bathing, organised) is used because it is the closest designated use to people entering water at a ghat. Values at the limit pass.
| Dissolved oxygen (DO) | Class B | 5 mg/l or more |
|---|---|---|
| BOD | Class B | 3 mg/l or less |
| pH | Class B | 6.5 to 8.5 |
| Dissolved oxygen (DO) | Class D | 4 mg/l or more (used only as a severity line) |
Class D DO line (4 mg/l) is used only to mark severity (the DO_SEVERE flag and the high-risk rule). It is not the standard a ghat is judged against.
Total coliform (Class B: 500 MPN/100 ml or less) is not in the data, so no ghat can be shown as meeting Class B in full.
Ranking and tiers (PRD v2 §6.4)
- Tier, using dissolved oxygen (Tier 1 = DO below 5 in more than half of readings; Tier 2 = any DO below 5; Tier 3 = no DO failure but BOD or pH outside limits; Tier 4 = no failure in readings taken).
- Share of DO readings below 5, highest first.
- Share of DO readings below 4, highest first.
- Share of BOD readings above 3, highest first.
- Median BOD, highest first; then site name A–Z.
DO leads because BOD is above the limit in more than half of readings at all 7 ghats (60% to 100%), so it barely separates them, while DO failure ranges from 0% to about 85%. Data gaps never lower a tier.
High risk = any DO reading below 4 mg/l, or DO below 5 mg/l in more than half of readings (PRD v2 §6.3).
Evidence states
Missing cells are never counted as passing. Each reading's Class B result is in the backend evidence CSV; these screens show printed values with criterion lines alongside.
- 6.4Reading. A printed number, including 0. Counted as a reading.
- BDL (below detection limit)BDL (below detection limit). Below detection limit: very low, not zero and not missing. Occurs only for metals here.
- ND (not detected)ND (not detected). Not detected: very low. Not present in the Bhopal file.
- Not sampledNot sampled. Blank, _, -, * or ** in the report. Missing; never counted as passing.
- No table in reportNo table in report. Expected cell with no row in the CSV (table or parameter absent from the report). Missing; never counted as passing.
- Unresolved flagUnresolved flag. Any other flag. Shown as printed with “meaning unresolved”.
- Not a possible readingNot a possible reading. A printed number that cannot be a real reading (negative). Reported as a data issue; never compared or counted as passing.
Expected coverage
1,638 expected cells (7 ghats × 6 seasons × 3 phases × 13 parameters); 1,524 rows in the file; 114 cells have no row at all.
114 expected cells have no table in the reports: no Durga 2025 table for Bairagarh (Sant Hirdaram) kund; no Durga 2025 table for Malikhedi kund; no suspended solids rows in Durga 2023 (7 ghats); no suspended solids rows in Durga 2025 (5 ghats). These cells are missing, not passing.
MPPCB visarjan reports 2023–2025 · Ganesh and Durga · before, during and one week after immersion · Track 2 data pack
What this tool cannot tell you
- No total coliform data, so no ghat can be shown as meeting Class B in full.
- No 2026 readings: everything is historical (2023–2025).
- Readings show levels, not causes. Pre-immersion BOD was already above 3 in 32 of 37 readings, so the data cannot show what immersion itself changed.
- No health effects are claimed. Readings are reported against CPCB limits only.
- Units are assumed; the reports print none.
- Exact sampling dates are not in the dataset; only year, festival and phase.
- Report page references come from the CSV. Only a partial manual spot check against the PDFs has been done; other page references are not yet checked.
- Only 4 metals were tested (chromium, lead, zinc, copper); they have no Class B limit and are reported, not judged.
- Recommended actions are unverified against CPCB idol-immersion guidelines. Hindi text is a provisional translation.
How to read this page
Tiers (PRD v2 §6.4)
- Tier 1DO below 5 in more than half of readings
- Tier 2Some DO readings below 5
- Tier 3DO met in every reading taken; BOD or pH outside limits
- Tier 4No reading outside limits among those taken. This is not a clearance.
Flags (PRD v2 §6.3)
- DO below 4Any dissolved-oxygen reading below 4 mg/l (Class D line).
- DO 0Dissolved oxygen read 0 mg/l.
- DO often below 5Dissolved oxygen below 5 mg/l in more than half of readings.
- BOD often above 3BOD above 3 mg/l in more than half of readings.
- pH outside rangeAny pH reading outside 6.5–8.5.
- Data gapAt least one DO, BOD or pH reading is missing. Missing is not treated as passing.
- Identity uncertainReport tables may describe two different places under this name.
Where each figure comes from
- Backend outputComputed by the Visarjan Ghat Watch Python backend from the supplied track CSV and CPCB criteria file; covered by its test suite.
- PRD v2 ruleA rule or plan-window description from the team's PRD v2 (text, not data).
- Not assignedIntentionally empty until an officer fills it in.