On-device planning model · no external AI service
This module scores every school and district with a small multi-criteria model computed in your browser from the
fused survey data (Census 2023, 2020 village survey, SED school list). New-school needs come from out-of-school
children vs current capacity. Upgrade recommendations score each primary/middle school on its enrollment
and the road-less distance to the nearest same-gender next-level school (girls' schools get a priority weight,
since travel distance is the strongest barrier to girls' continuation). Tick the recommendations you would
approve, then run the impact simulation — the model re-computes access distances as if your decision were
already implemented and reports the before/after difference. Distances are straight-line; real road distances
are typically 20–40% longer.
Priority logic. Every recommendation is classed
HIGH MEDIUM or
LOW by fixed thresholds:
settlements — High if no school within 8 km, the nearest is over 5 km away, or ≥500 school-age children are un-served (Medium: >3 km or ≥200 children);
hotspots — High at ≥2,000 unmet seats or ≥170% utilisation (Medium: ≥800 seats or ≥130%);
girls' gaps — High at ≥1,000 boys enrolled with zero girls' schools (Medium: ≥500);
upgrades — High at model score ≥70/100 (Medium: ≥55). Everything else is Low.