khalilurrahmanridoykhan/humanitarian-im-toolkit

Humanitarian IM & Rapid Mapping Toolkit — Data

Data from a humanitarian information-management and rapid-mapping toolkit: multi-sector rapid assessment (Kobo/ODK + Esri Survey123), administrative-boundary (P-code) integration, OSM-based rapid mapping, and a 4W/5W dashboard. Full source code, scripts and app: github.com/khalilurrrahmanridoykhan/humanitarian-im-toolkit.

No real beneficiary, camp-resident, or displaced-population data is used anywhere in this repo. No live KoboToolbox, ArcGIS Online, or OSM account was created or used at any phase — see the source repo's docs/ROADMAP.md.

Contents and provenance

data/ — synthetic site assessments + real admin boundaries

• —synthetic_site_assessments.csv — entirely fabricated (fixed random seed), does not describe any real site, event, or population.
• —cod/bgd_admin3_coxsbazar_subset.geojson — real, public data: a 6-feature subset of HDX's Bangladesh Subnational Administrative Boundaries (COD-AB), source Bangladesh Bureau of Statistics (BBS) via ITOS, distributed through UNOCHA ROAP, under OCHA's COD/FOD terms of use (see the HDX page for full terms). Only the boundary polygons are real; the synthetic sites placed within them are not.
• —upazila_pcode_crosswalk.csv — maps this repo's synthetic upazila slugs to official adm3_pcodes, including two intentionally-documented real name mismatches (ukhiya vs. official Ukhia, naikhongchhari vs. official Naikkhongchhari) that a name-based join would silently get wrong.
• —needs_choropleth_coxsbazar.geojson/.png, needs_choropleth.qml — synthetic site data aggregated to a composite needs-severity score per upazila, joined by P-code onto the real boundaries.
• —fictional_4w_activities.csv — fabricated 4W (who/what/where/when) activity data for the dashboard.
• —geohealth_integration/mock_hazard_placeholder.csv — values are explicitly marked SYNTHETIC_PLACEHOLDER; demonstrates a documented, file-based join contract to a companion project's hazard-risk layer, which doesn't exist yet at that project's current phase. Not real hazard data.

forms/ — the rapid assessment instrument

Multi-sector rapid site assessment XLSForm, IOM DTM/REACH-MSNA style: one enumerator visit produces one site-level record, never individual household/person data. Shipped in both Kobo/ODK and Esri Survey123 dialects (kobo-vs-survey123.md documents the concrete differences). The Protection section is explicit that no individually-identifying information is collected.

rapid_mapping/ — real OpenStreetMap extract + QA

Real, public OpenStreetMap data (693 building footprints, 71 road segments) for a small area in Ukhiya, Cox's Bazar — one of the most densely mapped places on Earth following Missing Maps/HOT campaigns after the 2017 Rohingya refugee influx. Fetched once via the public Overpass API and cached (cache/osm_raw_response.json). Includes a HOT-Tasking-Manager-style task grid and a QA/validation pass (38/693 buildings flagged as review candidates, not confirmed errors). © OpenStreetMap contributors, available under the [Open Database License (ODbL)](https://www.openstreetmap.org/copyright).

License

Apache-2.0 for this repo's own synthetic data, scripts, and forms. Real third-party data retains its own license/terms as noted above (HDX COD/FOD terms for admin boundaries; ODbL for OpenStreetMap data) — attribute accordingly if you reuse those specific files.