Project · Intermediate
WASH coverage and water quality by district
A district-level analysis placing every household on the JMP service ladders and testing whether chlorination is actually delivering safe water, so rehabilitation budget goes where the service is worst rather than where the sources look worst.
Context
- Organisation
- WASH programme covering three districts and eighteen communities
- Problem
- Coverage was reported by counting improved water sources, which ignored collection time entirely. Communities with a functioning borehole an hour away were counted as served, so rehabilitation budget went to the districts with the fewest boreholes rather than to the households with the worst service.
- Audience
- WASH coordinator and the district water authority
- Decision it informs
- Which six communities receive water point rehabilitation in the next funding cycle.
Datasets used
Deliverables
- Analysis notebook
notebooks/wash-ladders.ipynb - Dashboard
dashboard/index.html - Technical documentation
docs/indicator-definitions.md - Final report
reports/wash-coverage-review.pdf - Presentation slides
slides/district-briefing.pdf
Counting sources is not measuring service
The JMP drinking-water ladder does not classify a water point, it classifies a household. An improved source more than thirty minutes’ round trip is limited service, not basic — and in this survey about a quarter of households sit on limited service for that reason and no other. Every one of them is invisible to an analysis that groups by source type.
The project builds all three ladders — water, sanitation, hygiene — then reads them against the Sphere minimum standards of fifteen litres per person per day and a water point within thirty minutes.
The second question is whether the water is safe when it arrives
Chlorination coverage and water safety are different claims. This dataset tests free residual chlorine and E. coli on roughly a third of households, so the quality denominator is smaller than the access denominator, and reporting both against the full sample overstates how much is actually known.
The analysis has to state, for each figure, how many households it rests on. That is the habit the deliverable is really assessing.