Have I posted about that yet? If I have doesn’t matter, because that guy brings it on point!

Originally shared by Eli Fennell

Google Unveils Dataset Search for Researchers

The Internet Age has brought access to almost unfathomable volumes of data to the masses. For Researchers in the Sciences, Technologies, and other fields, this has been an even greater boon than it has been for most people. Countless hours in Research Libraries and handwriting correspondences to researchers and institutions, while not entirely eliminated, have been replaced with searchable indexes of Journals, Citations, and more, available anywhere that has internet access, with an increasing amount of these sorts of information becoming free to access and use.

Given the even more urgent necessity for good Search Algorithms to help modern Researchers, it isn’t surprising that Google, the Search Giants, long ago threw their own hat in the ring with Google Scholar, a free-to-access dedicated Search Engine for Researchers. Google Scholar also offers integration with Subscription Research Databases provided by Universities and other research Institutions, Organizations, and Facilities, to allow access to Paywalled Sources provided by these Databases to their users.

Now, they want to do much the same thing for an area which has, to-date, remained stubbornly arcane and old fashioned: Dataset Search.

Surprisingly (as I can personally attest), modern Researchers often find it quite difficult to access, and find relevant information within, the many publicly available datasets published by both public and private research agencies, publishers, institutions, and even individual researchers. This has always been the case, but the situation to-date has only slightly improved, even as other areas of research-related indexing and searching have taken leaps and bounds. Oftentimes, knowing how to find the right datasets can depend on word-of-mouth and fortuitous discovery, especially for cross-disciplinary research.

Google Dataset Search hopes to solve this problem, though surprisingly not (as yet) through the company’s traditional Algorithmic Magic. Instead, the initial offering relies on ranking and presenting Datasets by the use Schema.org web markup, which depends on publishers and hosts of data repositories to ensure that this is both done and done accurately. This Schema information allows those who ‘Own’ (originated/generated/published) the data to both ‘Claim’ it, so the Search Engine knows the Origin and Authority of the datasets, and to provide specific information about the nature and content of the datasets.

Depending on third-parties to make accurate and standardized web markup on all of their datasets in a timely manner would have been a Fool’s Errand for any company except Google, which was able to use its market dominance and reputation to cajole all of their Dataset Search Partner Universities to standardize their Schemas in under a year. Beyond Universities, their other Partner Agencies have also been stepping up quickly to get onboard with this effort, as well. This will help Google both rank Datasets in Search, and integrate their Knowledge Panel into these results.

In the future, they may also experiment with Algorithms, Machine Learning, and other techniques for ranking and displaying dataset results, but for now the marriage of Schema markup and the Google Search brand name and power are uniquely well suited to the task of organizing and surfacing the right datasets from the countless open sources now available online.

#BlindMeWithScience #OpenAccess #DatasetSearch

https://www.nature.com/articles/d41586-018-06201-x?WT.feed_name=subjects_computational-biology-and-bioinformatics//cdn.embedly.com/widgets/platform.js