n x m

Here’s how it works:

+ Take the m most relevant photos from Flickr matching the keyword n.
+ Aggregate all keywords for these m photos and sort them by frequency.
+ Find the most popular photo matching each aggregated keyword with at least 2 occurrences as well as n.

Notice that the Lusaka one contains no less than 13 photos of me, and some more by me.

Related:  Dopplr.org