Dataset Search

2020: Start

In January, 2020 Google started service for convenient search of data sets of Dataset Search after its testing for more than one year.

This tool allows to find on a key word the necessary information among about 25  million repositories freely available in  the Internet. Sorting of search results according to date of updating, rights to use, costs of access and  to a format of downloading is available.

Google started service for convenient search of data sets

According to the SiliconANGLE edition, open data sets play more and more important role in the technology sector against the background of fast distribution of artificial intelligence. The AI is more difficult, the more training data need to be processed that the algorithm became ready for serial use. Similar Dataset Search portals where developers of AI systems can look for on a centralized basis to data, can become the valuable tool for projects of machine learning.

Besides, the data sets usually  provide  research laboratories and  the large organizations,  such as  Facebook.

Before full-scale start of Dataset Search the Google company significantly finished service in comparison with the beta. In particular, "the quality of descriptions for repositories was considerably improved" and the new filters allowing users to narrow search results depending on what data set is required to them are added.

Now you can filter results on the basis of desirable types of data sets which you want (for example, tables, images, the text), or on the basis of whether the data set at the supplier is available. If the data set is connected with geographical area, you can see the card — the researcher of Google Natasha Noy wrote in the blog.

It is noted that developers of service aimed at  the standardized description of the data  offered  by the schema.org^[1]

Notes