Service demand is not water quality.
A modeled relative indication of service demand. It does not measure household water quality or drinking-water safety.
A neighborhood score cannot identify a contaminant, diagnose a household water issue or certify water as safe to drink. No modeled values are published until actual source-backed inputs exist.
Current methodology
Version 1.0 · Published 2026-09-06. Methodology changes create new versions and snapshots. Historical scores are preserved.
| Component | Configured weight |
|---|---|
| Search demand | 30% |
| Seasonal / climate context | 10% |
| Property / accommodation density | 20% |
| Anonymized request activity | 15% |
| Provider-gap signal | 10% |
| Public water context | 15% |
Normalization and missing values
Component inputs are stored separately on a 0–100 scale. A published score is the weighted mean of available components: sum(value × weight) / sum(available weights). Raw imported values are retained with their transformations; normalized snapshots do not overwrite them.
Renormalize weights across available source-backed normalized components. Publish no score when no components exist. Never impute neighborhood values from provincial statistics.
Privacy comes before a chart
Public trends require at least 20 independent valid requests per cell by default. The hard minimum is 10. Small cells and raw comments are withheld. Exact consumer addresses, coordinates and contact details are never public map layers.
Provider coverage
Coverage counts use the live registry of active, verified participating providers with accepted terms and available capacity. They describe platform participation, not all businesses in an area. Actual routing also checks service fit, caps and lead permissions.
Geography and dates
Consumer areas use approximate curated point locations. Natural Earth land geometry is geographic context, not a neighborhood boundary. Province-level information remains province-level. Every published data point retains source geography, period, retrieval date and methodology.
Review cadence
Editorial guidance is reviewed every 6–12 months. Data snapshots are reviewed monthly or when meaningful new source inputs are available. Utility notices require their own dates and expiry. Corrections can be requested through contact.
Source registry
Centers for Disease Control and Prevention · General travel guidance; not a Bali household assessment · Retrieved 2026-09-07
Linked and summarized; CDC source. No local water-quality measurements inferred.
Centers for Disease Control and Prevention · General treatment principles; not Bali-specific · Retrieved 2026-09-07
Linked and summarized; verify product-specific claims independently.
US Geological Survey · General water science; not a Bali hardness survey · Retrieved 2026-09-07
US government educational reference; attribution retained.
US Environmental Protection Agency · US private-well guidance; local requirements must be confirmed in Bali · Retrieved 2026-09-07
US government guidance, linked as general context; not Indonesian regulation.
Centers for Disease Control and Prevention · General filter-selection guidance; not a Bali product endorsement · Retrieved 2026-09-07
Linked primary source; product-specific claims require verification.
US Environmental Protection Agency · US aesthetic water guidance; not Indonesian limits or local measurements · Retrieved 2026-09-07
General educational context only; no US threshold is presented as Indonesian law.