I apologize for breaking the rules and making this post a bit geeky. I noticed the geotagging feature in Flickr and couldn't help writing about it. Ok, so lets quickly get to the point here.
I find the graph generated by geotagging quite similar to the web graph. To illustrate this analogy, lets take an example. Say, a person living in NY has geotagged her vacation photographs for Paris and Sydney. Consider the cities as nodes and draw the weight 1 edges NY->Paris and NY->Sydney. Accumulate these edges across users and construct the final graph that might look like the figure below.
The big circles correspond to big sites like Yahoo! with heavy inter-linkage between its sub-domains and main page. The analogy would be to consider a major tourist city like Paris with other smaller cities around it. Heavy inter-linkage with nodes around you means that people often visit nearby places to chill out for short durations. I guess it should be easy to identify such neighbourhoods based on their coordinates.
Now if a place is a major tourist attraction, it is bound to have a lot of incoming tourists from all over the world just like popular sites have lot of incoming links. The Page Rank algorithm should give us a nice ranking score for tourist destinations. Looking from a random surfer perspective, a person on a random world tour taking random tourist flights would be more likely to spend a large amount of time on highly popular places.
Further, if we know the attributes of a place like is it famous for beaches, mountains, city life, etc., these attributes could correspond to topics and we can generate topic sensitive rankings.
Well, it could be just a random thought of mine or it could be something. May be, its very trivial and/or they already do something like this. I have no idea. Sounds like fun to me. What say ?
No comments:
Post a Comment