TGISAI Readiness for Transport Data
Making the world's transport data queryable by AI agents — and by anyone who can ask a question in plain language.
The problem
Global transport data is scattered across 25+ platforms — World Bank indicators, Oxford's OPSIS infrastructure maps, IMF PortWatch trade disruptions, OECD transport statistics, ADB's Asian Transport Outlook, the Africa Transport Observatory, and more. Each platform has its own API, its own query language, its own conventions. They were built for specialist users who know where to look and how to extract what they need.
Efforts to align this fragmentation are underway. The Transport Data Commons (TDC) is building the shared catalogue — a single place to discover what transport data exists globally, with standardised metadata and open access. That alignment work is essential. But even when data is catalogued, it still isn't ready for AI agents to query directly. APIs differ, schemas clash, geospatial and tabular data live in different worlds, and no agent can move between them without source-specific guidance.
That's the gap TGIS fills. TDC aligns the data. TGIS makes it AI-ready.
What TGIS does
TGIS is a service layer that sits alongside TDC and other platforms — complementing them, not replacing them — and makes their data discoverable and queryable by AI agents. A user asks a question in natural language. The system works out which datasets are relevant, queries them, combines the results, and returns an answer.
Several organisations are building AI chatbots for transport data — ADB is exploring one, TDC is developing another, CKAN-based bots are appearing on open data portals. Each talks to a single organisation's data. TGIS provides the cross-source layer underneath: the connective tissue that any of these chatbots could plug into to query across multiple sources in a single flow. It makes TDC's chatbot better, not competes with it.
Three things make this work:
- Discovery — a structured registry that tells AI agents what data exists, what questions each source can answer, and how to access it
- Comprehension — machine-readable descriptions of schemas, fields, and query patterns so agents can use APIs they've never seen before
- Access — lightweight wrappers where needed (some APIs are already agent-friendly; others, like ArcGIS or SDMX, need translation)
Who's behind it
TGIS is led by the RIDE programme (Research on Infrastructure in Developing Economies), funded by UK International Development / FCDO, with technical delivery supported by the Frontier Tech Hub. The aim is a system built in partnership with multilateral development banks (ADB, AfDB, EBRD, EIB, CAF, IDB), UN agencies (UNECE, UNDESA), and sector organisations — not owned by any single institution.
TDC (Transport Data Commons) is the bedrock. TDC is the shared catalogue of what transport data exists globally — the data commons the community maintains. TGIS adds an AI layer on top that makes TDC's datasets, and data from other platforms, queryable by any agent. If TDC is the library catalogue, TGIS is the librarian that can read across all the books.
Why now
The UN Decade of Sustainable Transport (2026–2035) launched in December 2025 with monitoring goals but no data infrastructure to track progress. AI capabilities have reached the point where querying heterogeneous data in natural language is technically feasible and affordable. And the data already exists — our audit found 14 of 25 sources with open APIs requiring no authentication. The bottleneck is fragmentation, not availability.
Starting point
A proof of concept with 10 sources spanning different data types, transport modes, and institutions:
| Source | What it covers |
|---|---|
| World Bank Open Data | Country-level transport indicators |
| OPSIS (Oxford) | Global infrastructure networks + climate risk |
| PortWatch (IMF/Oxford) | Maritime trade disruptions |
| OECD / ITF | Transport statistics across member countries |
| IATI / DAC | Aid and investment flows into transport |
| ADB Asian Transport Outlook | 450+ indicators across 51 Asia-Pacific economies |
| Africa Transport Observatory | Continental transport corridor data |
| TDC | Multi-modal datasets via single API |
| AfDB Data Portal | African infrastructure and economic indicators |
| SLOCAT NDC Tracker | Climate policy commitments mapped to transport |
This is deliberately not dominated by any single institution. The full data registry covers all 25+ audited sources, with active wrappers for these 10.
What it enables
With these sources connected, an AI agent can answer questions that no single platform handles today — combining infrastructure condition with spending data, overlaying climate risk on trade routes, tracking policy commitments against actual investment. The principle: data that was never designed to be connected, made connectable through AI.
TGIS supports queries in any language, works at both global and country level, and serves any AI tool — not just a single chatbot.