Use data from different Semantic Mediawikis and also Communecter's database


(Simon Sarazin) #1

I’m actually starting to put data of “commons” projects on different Semantic Mediawiki :

  • transformap
  • openlaw.fr (list of juridical commons, they just started)
  • wiki.lafabriquedesmobilites.fr (list of commons in the mobiliy sector, i’m working on this mediawiki and starting to get data of different commons projects)
  • movilab.org (lot’s of data on open coworking/third places in France, not in semantic for the moment)

We have also more and more data on communecter.org platform (based on a centralised database that we call the “opencommondatabaseXX”). For example here is a starting list of third places projects in France ( https://www.communecter.org/#search:#tierslieux:all) , that is also accessible in parts on the movilab.org mediawiki. The same for the commons https://www.communecter.org/#search:#commun:all
How could we connect the data, and not duplicate it between different mediawiki using semantic extension and also between mediawiki and communecter ? I would prefer to use the transformap data than to create a new one in communecter or in an other semantic mediawiki

Any direction to find a solution ? Things to try to do ? Connectors to developp ?

Thanks :slight_smile:


(Jon Richter) #2

Your questions lead into the right direction. Unfortunately my head is preoccupied with few technical considerations around deduplication, which makes it harder for me to answer them. I believe @species has an opinion on how to do this?

What you are basically asking to me is: Do we have a way to seamlessly integrate linked data from various sources, e.g. how DBpedia does, or will we keep synchronised datasets from each other.

Then you are talking about what @otisyves has been building in the Remix the Commons wiki. See their article about http://wiki.remixthecommons.org/index.php/Commons#EN for example, which links to Wikidata, Wikipedia, DBpedia and P2P Foundation sources. Therefore the Remix the Commons wiki is a suitable place to place further connections, turning into a specialised centrality within the wealth of data sources available within the network.

I would propose we further dive into

and start adding multiple of such links to our multiple data sources into one of the Semantic MediaWikis at hands. Thus, we’d be building a pretty classical federation.

Then we could start building a crawling client and inference machine which is able to combine the data of those sources into one and display it.


(Jon Richter) #3

@toka In following up on our recent phone call, this is where I had described a little context about how to strategise mixing the contents of multiple Semantic MediaWikis. Any comments or recommendations welcome.