Discover millions of datasets on Google Dataset Search

Vlad Tămaș
Accesa
Published in
2 min readJan 27, 2020

Google has launched a search service for machine learning datasets. Google Dataset Search allows you to search data from 25 million open datasets. The search system differs from analogues in that Dataset Search does not collect them manually and does not store datasets at home. Dataset Search indexes open datasets using a standard data structure description format.

New features in Dataset Search

Users can now filter search results based on the type of dataset and public availability of the dataset. The type of dataset is determined by the type of data it contains: images, text, or tabular data. In addition, the system is now adapted for mobile devices. The creators made the dataset descriptions more complete. To publish a dataset to Dataset Search, a user must describe its structure using standard Schema.org .

Dataset Search targets a wide audience: researchers, students, business analysts and date Scientists. The most popular queries for the system are education, weather, cancer, crime, soccer and dogs.

What datasets are contained in the search engine

Dataset Search provides insight into open data on the web. Topics covered by the data are geosciences, biology, and agriculture. Most governments publish their data and describe it with Schema.org. The United States leads in the number of open government databases. Their number totals more than two million. The most popular data format is tables — over 6 million in Dataset Search.

Source: https://blog.google/products/search/discovering-millions-datasets-web/

--

--

Vlad Tămaș
Accesa
Writer for

AI Consultant focusing on bringing intelligence into our customers next generation on products.