UNDP RAPIDA

Fri, 01 Nov 2024 00:00:00 +0900

Period: November 2024 – August 2025

Organization: United Nations Development Programme (UNDP)

Links: GitHub (UNDP-Data/rapida) (BSD-3-Clause)

Overview

RAPIDA is a semi-automated geospatial analysis tool for rapid crisis response. When a crisis such as a natural disaster happens, it quickly assesses the exposure of people, assets and the economy within an area of interest by computing zonal statistics from curated global geospatial datasets. The results can be shared through UNDP GeoHub so that decision makers can use them right away.

Features

Fetches and curates administrative boundaries from OpenStreetMap and OCHA
Assesses multiple exposure components over an area of interest:
- Population: WorldPop (constrained, UN-adjusted)
- Buildings: merged Google, Microsoft and OpenStreetMap building footprints
- Electrical grid: predictive global distribution networks
- Roads: GRIP global roads dataset
- Land use: Sentinel-2 imagery with the Google Dynamic World model
- Socioeconomic: deprivation index, relative wealth index and GDP
Project management on UNDP Azure storage and publishing to GeoHub
Integration with JupyterHub for notebooks and visualization
Rich command line interface: init, auth, admin, create, assess, list, download, upload, publish, delete

Tech stack

Python, GDAL, rasterio, exactextract, GeoPandas, Click, Rich, Azure Blob Storage, JupyterHub

UNDP GeoHub

Wed, 01 Dec 2021 00:00:00 +0900

Period: December 2021 – August 2025

Organization: United Nations Development Programme (UNDP), SDG Integration Team, Bureau for Policy and Programme Support

Role: GIS Consultant → Full Stack GIS Developer

Links: GeoHub · GitHub (UNDP-Data/geohub)

Background

UNDP has developed a corporate data strategy to accelerate its transformation into a data-driven organization, and geospatial data plays an important role in it. However, the large-scale adoption of geospatial data had been obstructed by data silos across country offices, lack of interoperability, and limited infrastructure and know-how. GeoHub was developed as a part of the Data Futures Exchange (DFx) to solve these issues.

What is GeoHub?

GeoHub is a centralized ecosystem of geospatial data and services to support development work. It allows users without geospatial skills — UNDP staff and policymakers — to:

search, visualize and download a wide array of datasets, including satellite imagery and spatio-temporal model data
upload their own datasets through a data upload pipeline
compute dynamic statistics and run raster analytics
create maps and share them with the community easily

Architecture

Frontend: SvelteKit and MapLibre GL JS web applications
Backend: PostgreSQL/PostGIS, titiler, pg_tileserv and other FOSS4G software deployed on Azure Kubernetes Service (AKS)
Cloud: Azure App Service, Blob Storage, PubSub
Data: Cloud Optimized GeoTIFF, vector tiles and STAC catalogues
All source code is published on GitHub under an open-source license

Main features I developed

Landing page to direct users to each part of GeoHub
Social login (GitHub authentication and UN B2C authentication)
Data upload pipeline
User permission management for datasets and maps
UI/UX improvements for data visualization with MapLibre
Raster analytical features (titiler algorithms)
Static image API to generate static map images from a saved MapLibre style

See also RAPIDA, a crisis assessment tool integrated with GeoHub.

Skills

MapLibre, Mapbox, Svelte/SvelteKit, Node.js (JavaScript/TypeScript), Python, PostgreSQL/PostGIS, titiler, pg_tileserv, STAC, Azure (AKS, App Service, Blob Storage, PubSub)

UNDP | JIN IGARASHI