Linked, open data and Knowledge Graphs show their full power when they are connected. In technical terms this is called federation. A query across multiple data sources is then a federated query.
For example, an item from FactGrid is linked to the corresponding item in Wikidata to retrieve complementary information. This way, there is no need for redundant data in two different data sources, which in case of doubt are not synchronized.
A very simple query shows the partners of Magnus Hirschfeld, a renowned sexologist in the 1920s, from Wikidata as well as DBpedia, a knowledge graph derived from Wikipedia. It shows: Both data sources have partners, but different ones and both are correct. Only a federated query gives the full picture. Unfortunately, we cannot join FactGrid data. Because as of now, Wikidata only allows a “selected number of other SPARQL endpoints” for this type of decentralized queries. If FactGrid wants to participate, we have to get in line. FactGrid has now done that, we have made a nomination for ourselves.
Theoretically, all we need are external identifiers, for example to Wikidata or Wikipedia articles or other data sets such as the GND. A lot of FactGrid items have this information stored anyway. Perfect conditions to become part of the distributed ecosystem.
How long does this process take? No idea.
Michael Ringaard’s KnolBase, a prototype wikibase browser, gives a taste of the potentials. Based on the daily data dumps of FactGrid and Wikidata, KnolBase accumulates information from both sides. The Wikidata-Identifiers on FactGrid allow Ringaard’s browser to basically understand where the same has been said on both sides and where the information essentially differs. The result is no longer a side by side presentation of all the results from different pages but a new uniform page that intelligently presents all the information it has compiled – like a human reader would do after collecting and comparing the information of various sources. This is the KnolBase page on Adam Weishaupt and it is more than Wikidata or FactGrid offer on him:
A background to Wikidata federation and future plans is provided by Bayan Hills, who works at Wikimedia Germany, in her talk at the ld4 conference on linked data 2021, starting at 19:00:
Another good talk on querying on a decentralized web: