UI/UX & Frontend for the World Bank Metadata Editor

Statistical metadata is dense, deeply structured and governed by a different international standard for every data type — DDI Codebook for surveys, ISO 19139 for geospatial data, Dublin Core for documents, IPTC for multimedia — yet the people entering it are statisticians, not interface specialists. The editor had to make long technical forms navigable, adapt its interface to seven distinct data types, handle real data file parsing, and slot into the World Bank's existing publishing ecosystem (IHSN tools, Nesstar outputs, NADA catalogs) — all while an aging codebase was being replaced underneath it.

We designed and built a dynamic form-based editing interface organized in sections and subsections that mirror each metadata schema: DDI Codebook 2.5 for survey microdata, SDMX-style structures for indicators and time series, ISO 19139 for geospatial datasets, Dublin Core for documents, IPTC for images and video, plus statistical tables and reproducible research scripts. Documentation work is organized in projects grouped into nested collections, with faceted search across type, ownership and collection trees, and granular sharing roles (viewer, editor, administrator). A statistical file parser imports CSV, Stata and SPSS files and extracts variable lists, labels and counts automatically, feeding an inline variables editor with dynamic variable groups. Metadata imports and exports flow through DDI XML and JSON for continuity with existing IHSN tooling, and a publishing workflow pushes finalized packages straight to NADA data catalogs over REST. The administration portal covers analytics, user and role management (Laravel Permission RBAC, Sanctum tokens, email OTP for admins), controlled vocabularies and a translations manager. The stack ships in private GitHub repositories with Docker deployments to World Bank cloud environments.