Thanks, these are incredibly useful pointers. More comments follow...
I have the precise same problem of having photographs and trying to
extract meaning from the clusters. I've been working on code to
scratch this itch, and I'd be happy to send it to anyone, or to work
with someone else to generalize the solution. The code is in Perl.
I haven't worked in Perl for a few years (do more PHP and Python now) but I'd be very interested in seeing your code—and to help with the project as well, though I don't know how much help I could be :~(
I also have track logs for all of the points where the photos were
taken. So I know when I was near to each photo (both when it was
initially taken, plus subsequent visits to the area). I also have a
collection of waypoints for the general area. And finally, I have a
collection of travel ephemera like ticket stubs and receipts, that all
have time stamps and which I've been geocoding based on the track
logs.
My project sounds somewhat similar to yours in its goals, except that instead of your data riches (tracklogs, waypoints, etc.), I'm looking to bootstrap from a scanty collection of data, in order to incent the creation of more data.
Here's the background: Since Flickr introduced their cumbersome but oh-so-broadly accessible geotagging interface, I've been thinking about the problem of how to incent more non-geowanking civilians to geotag their photos, and ultimately to be interested in having the means to geotag automatically, rather than being more fearful than hopeful about that prospect, as many seem to be. One answer I came up with is writing a webapp for Flickr users, a sort of geo-customized startpage that can deliver local info/recommendations uniquely interesting to you based on a) the geographic area where you have taken and geotagged the most photos and b) the other metadata that exists for those photos: times, tags, titles, comments, etc. Ideally, the app would produce at least slightly relevant results if the user geotagged even a small part of their photo collection, and (like recommender systems), the more datapoints were added, the better-fit the content it could serve. (I'm taking an applied NLP class this semester so one thing I'm hoping to look at is the potential of geo-context + NLP to support these kinds of inferences.)
The pictures where 'Vodensky' was the nearest waypoint were, sure
enough, best tagged as 'Vodensky' for the Vodensky Military Museum.
And the pictures closest to the waypoint 'ASIEN-GIRLS' were in fact of
the strip club that featured 'Asien Girls.'
Interesting. I'd love to try doing localized web searches on your waypoint names to see if there's any trend in what sort of results come back. (Though in the case of 'Asien Girls,' I can pretty well imagine...)
Schuyler wrote some cool clustering code for Google Maps Hacks. Here
is an example of the code in action:
http://mappinghacks.com/projects/gmaps/cluster.html
It looks like this code would be a great starting point for me at the least. I will finally pick up a copy of GMH and look into it.
The gazeteer lookup is either fairly easy, or a pita.
Yeah, it seems so ;). Thanks for the pointer to GNIS. It would be great if someone could come with a way of using folksonomies + gazeteers to improve geo-naming, esp. at the region or neighborhood level. Do you know if that's something that either these Neighborhood Project people or the folks at Metacarta are thinking about?
I've been working on some code to implement a geographical data store
that would sort of intrinsically allow for the creation of user
defined 'areas' or 'regions' or 'neighborhoods.'
Cool! What would be the interface for defining them?
Thanks tons,
--Andrea
_______________________________________________ Geowanking mailing list [email protected] http://lists.burri.to/mailman/listinfo/geowanking
