12 GB of metadata and 50 TB of files released by Yahoo!. «Plus about 49 million of the photos are geotagged!» http://yahoolabs.tumblr.com/post/89783581601/one-hundred-million-creative-co...
Can this dataset help linking projects' articles and Wikidata items to the images we miss? I can imagine * a FIST expansion querying the data in some way to find flickr images "nearby" a geocoded page of ours, * a bot mapping photo IDs from geolocated Wikidata entries and then a bot importing on Commons those we lack, * a Wikidata Game to aid automation in some of the above. Where an image is in a non-free Creative Commons license, we can flickrmail the author to relicense it, the success rate is typically high (we could also do this by bot, if we're able to automatically write a message mentioning the specific files and the pages where we'd use them).
Nemo
If the metadata dataset license allow it, it would be interesting to import it into Labs as a database.
2014-07-05 8:47 GMT+02:00 Federico Leva (Nemo) nemowiki@gmail.com:
12 GB of metadata and 50 TB of files released by Yahoo!. «Plus about 49 million of the photos are geotagged!»
http://yahoolabs.tumblr.com/post/89783581601/one-hundred-million-creative-co...
Can this dataset help linking projects' articles and Wikidata items to the images we miss? I can imagine
- a FIST expansion querying the data in some way to find flickr images
"nearby" a geocoded page of ours,
- a bot mapping photo IDs from geolocated Wikidata entries and then a
bot importing on Commons those we lack,
- a Wikidata Game to aid automation in some of the above.
Where an image is in a non-free Creative Commons license, we can flickrmail the author to relicense it, the success rate is typically high (we could also do this by bot, if we're able to automatically write a message mentioning the specific files and the pages where we'd use them).
Nemo
Commons-l mailing list Commons-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/commons-l
Emilio J. Rodríguez-Posada, 05/07/2014 11:00:
If the metadata dataset license allow it, it would be interesting to import it into Labs as a database.
I'm not sure they're considering it copyright eligible, but one just has to ask: http://webscope.sandbox.yahoo.com/catalog.php?datatype=i&did=67 Where copyright eligible, as they say it "is compiled from data available on Yahoo! Flickr", the metadata of -SA images would be in the same license.
Nemo