Data source coverage · How the retrieval engine works

100+ authoritative data sources across 11 categories

For each search, the system activates the source groups relevant to your topic and region, judges every candidate item by item, and attaches a traceable source to each result. This page lists the sources the retrieval engine currently connects to, so you can see the coverage before you submit a request.

91 mainstream dataset catalogs, grouped into 11 categories

4 targeted sources searched site by site

5 platforms connected by direct API, returning the platform's real metadata

11 source categories, activated on demand by topic

HOW RETRIEVAL WORKS

Sources are activated on demand, not carpet-searched

The 100+ sources are not all queried on every search. The system first understands your request, then picks the relevant source groups and runs a real search against them — so results stay focused and every source is checkable.

① Match groups by topic and region

A medical request activates the medical and life-sciences group; a request for Chinese provincial and city data activates the Chinese-platform group. Relevant groups come first, and unrelated ones don't waste a search path.

② Run a real search on the activated sources

For platforms connected by direct API, we return the platform's real metadata — dataset name, license, size and more; for catalog-type sources, we run a targeted search of their public data pages on site.

③ Judge item by item, with a traceable source

Each candidate is judged against your must-have conditions one by one, showing hits and gaps, and every result carries a source link so you can open the original page and verify it yourself.

SOURCE LIST · 11 CATEGORIES

91 mainstream dataset catalogs, grouped by discipline and region

Below are the catalog-type sources the retrieval engine currently connects to, matching the list built into the code. The list keeps expanding; what's connected at search time is authoritative.

Chinese dataset and competition platforms19 sources

Leading Chinese dataset platforms and provincial and city public-data portals, activated first for Chinese-language requests and China regional data.

Alibaba Tianchi datasetsBaidu AI Studio datasetsDataCastleOpenDataLabModelScope datasetsOpenXLab datasetsScience Data BankNational Earth System Science Data CenterYujian DatasetsHyper.ai datasetsOpenI datasetsLUGE (Qianyan) datasetsBeijing Public Data Open PlatformShanghai Public Data Open PlatformGuangdong Open Data PlatformShenzhen Open Data PlatformZhejiang Open Data PlatformShandong Public Data Open NetworkHangzhou Open Data Platform

Regional and non-English government open-data catalogs15 sources

Government open-data portals across Asia, Latin America, the Middle East and Africa, activated when a request concerns a specific country or region.

Singapore Open Data data.gov.sgJapan Data Catalog e-GovHong Kong SAR Open Data DATA.GOV.HKTaiwan Open Data data.gov.twMacao SAR Open Data data.gov.moKorea Public Data Portal DATA.go.krIndia Open Data data.gov.inMalaysia Open Data data.gov.myThailand Open Data data.go.thIndonesia Open Data data.go.idBrazil Open Data dados.gov.brMexico Open Data datos.gob.mxChile Open Data datos.gob.clSaudi Open DataOpen Africa

International-organization statistics catalogs9 sources

Cross-country comparable macro, trade, health and population statistics — the go-to benchmarks for international comparative research.

IMF DataUN DataWHO DataFAOSTATEurostatUN ComtradeILOSTATUNICEF DataWorldPop

Earth science and remote-sensing catalogs9 sources

Authoritative publishers of weather and climate, satellite remote-sensing and geospatial data, activated for geoscience and environmental requests.

NOAA DataNOAA NCEIUSGS ScienceBaseUSGS EarthExplorerCopernicus Data SpaceNASA Earthdata SearchMicrosoft Planetary ComputerNatural EarthGeofabrik OSM downloads

Medical and life-sciences open-data catalogs7 sources

Public databases for physiological signals, genome sequencing, neuroimaging and clinical trials, activated for medical and life-sciences requests.

PhysioNetOpenNeuroNCBI SRANCBI BioProjectEMBL-EBI BioStudiesEMBL-EBI Expression AtlasClinicalTrials.gov

Machine-learning and paper dataset catalogs5 sources

Dataset catalogs for model training and paper reproduction, activated for machine-learning and AI requests.

OpenMLUCI Machine Learning RepositoryPapers with Code datasetsRoboflow UniverseTensorFlow Datasets

Cloud-provider and large-scale open-data catalogs5 sources

Large public datasets and web corpora hosted on cloud platforms, activated for large-scale corpus and compute-side requests.

AWS Registry of Open DataGoogle Cloud public datasetsAzure Open DatasetsMicrosoft Research Open DataCommon Crawl

Academic data repositories and repository registries4 sources

Research data-deposit repositories and repository registries, activated for paper-supplement data and cross-repository lookups.

Mendeley DataOpen Science Framework (OSF)OpenAIREre3data registry of research data repositories

Social science and microdata survey catalogs4 sources

Authoritative archives of household surveys, panels and population microdata, activated for social-science and population requests.

ICPSR social science data archiveUK Data ServiceGESIS data catalog (Germany)IPUMS population microdata

Macro-finance and company disclosure catalogs4 sources

Macroeconomic series, census data and public disclosures of listed companies, activated for economics and finance requests.

FRED (Federal Reserve Economic Data)U.S. Census Bureau dataSEC EDGARNasdaq Data Link

General dataset catalogs2 sources

Cross-disciplinary general dataset search entry points, usable as a supplementary path for any request.

Google Dataset SearchData.world

Targeted sources + direct API4 + 5 sources

Targeted sources are searched site by site for public data pages; direct-API platforms return the platform's real metadata — dataset name, license, size, download count and more.

Targeted sources

National Bureau of Statistics of ChinaCnOpenDataHeywhaleOECD

Direct-API platforms

HuggingFaceKaggleZenodoFigshareHarvard Dataverse

The source list keeps expanding

This list stays in step with what the retrieval engine actually connects to, and grows and changes as new sources come online. What's connected at search time is authoritative.

Every result carries a traceable source

Each result comes with a source link, so you can open the original page and verify it yourself. We only point to real sources — we don't re-host or redistribute the original data.

Misses are reported honestly

Broad coverage doesn't guarantee a hit. When no matching data is found, the report states the gap honestly and explains the search direction and any near-match sources.

Describe your data need and we'll run a real search across these sources

New users get 3 free points on sign-up, enough for 3 standard availability checks. When a check finishes, you receive an availability report with hits and gaps explained item by item.

Start an availability check Explore research services