tag:blogger.com,1999:blog-6569681.post1279929330899140827..comments2024-03-24T10:38:16.997-07:00Comments on Geeking with Greg: Travel itineraries from Flickr photo trailsGreg Lindenhttp://www.blogger.com/profile/09216403000599463072noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-6569681.post-38947170728626388162010-12-07T14:59:39.635-08:002010-12-07T14:59:39.635-08:00I was very impressed to learn the student was able...I was very impressed to learn the student was able to actively collaborate with the Flickr team at Yahoo n this research. Do you think this might result in a real product that they could bring to the photo sharing service?MattHursthttps://www.blogger.com/profile/15951788258339477876noreply@blogger.comtag:blogger.com,1999:blog-6569681.post-69657442797754973762010-06-09T14:26:45.035-07:002010-06-09T14:26:45.035-07:00That WWW2010 poster does use Flickr photos, filter...That WWW2010 poster does use Flickr photos, filtering for ones that have GPS tags on them, but only considers each photo in isolation; that poster does not generate timestamped trails of movements, duration of stays, and transition times.<br /><br />What I think is the very clever idea in the HT2010 paper is that you can build trails of where people have been, linking each photo together to form a path, by combining the timestamp and location (either explict or derived from tags) data. It's the breadcrumbs, the trails, the history that is new, clever, and apparently quite powerful.Greg Lindenhttps://www.blogger.com/profile/09216403000599463072noreply@blogger.comtag:blogger.com,1999:blog-6569681.post-73261290456172495572010-06-09T14:01:28.839-07:002010-06-09T14:01:28.839-07:00A related poster was also published at WWW'201...A related poster was also published at WWW'2010: <a href="http://domino.mpi-inf.mpg.de/intranet/ag5/ag5publ.nsf/e127ff338913b2a3c12565f4005ef860/249c12c7ebc7712fc125773d00733687/$FILE/p1121-jain.pdf" rel="nofollow">Antourage: Mining Distance-Constrained Trips from Flickr</a>.Stepsnoreply@blogger.comtag:blogger.com,1999:blog-6569681.post-85779096004905605942010-06-09T07:06:23.278-07:002010-06-09T07:06:23.278-07:00Hi, Marin, The big difference is the use of billio...Hi, Marin, The big difference is the use of billions of Flickr photos, which is the interesting idea in this paper. The paper you cited used blogs as the data source. <br /><br />The thing to note in this HT2010 paper is the very clever idea of how Flickr photos can produce timestamped trails of movements, like coarse-grained GPS trails.Greg Lindenhttps://www.blogger.com/profile/09216403000599463072noreply@blogger.comtag:blogger.com,1999:blog-6569681.post-86820492931634754982010-06-09T04:24:56.550-07:002010-06-09T04:24:56.550-07:00there was a similar paper from Microsoft at WWW...there was a similar paper from Microsoft at WWW'2010 - <a href="http://research.microsoft.com/pubs/121309/traveloguemining_www2010.pdf" rel="nofollow">Equip Tourists with Knowledge Mined from Travelogues</a>Marin Dimitrovhttps://www.blogger.com/profile/15913353286682544452noreply@blogger.com