in Libraries/Info Sci, Ramblings, Tech

The Unsearchables

I’ve been having a lot of fun playing Colorwar on Twitter lately. The most recent event is a scavenger hunt run in Google Street View. GSV is made up of pictures taken from a car driving around various cities, and you can pretend you’re driving around as well. It comes in handy for getting a view of what a storefront looks like before going there, and similar simple tasks. But it also inadvertently captured people on the street and random events around town. Finding some of these events has been a real challenge for the scavenger hunt. For example, we needed to find a crossing guard. The presence of one relies on sheer chance that school was getting out just as a GSV truck drove by. Pretty slim odds, really. But there is absolutely no way to search for something like this other than simply checking every school in existence.

Some third party sites have stepped in and tried to fill that gap a little. Google Street View Gallery allows users to post links to specific images, which they can then tag. The tags and descriptions are then searchable. This system is handy, but still relies on a user randomly stumbling across something, thinking it is interesting, and then taking the effort to post it to the gallery with an adequate description. A very long chain of events, and something most people won’t deal with.

Google Street View represents a massive database of public images. All privacy issues aside, I would love for Google to test out some kind of image-recognition search in it. That would reveal a huge wealth of usable data. Or, with Google Maps’ APIs, maybe it’s possible for someone else to work on this. I don’t know.

But what I do know for sure is that we’ll only see more massive chunks of image data as time goes by. There will simply be too much for human eyes and minds to process – organizing it all will be a big challenge, and not something I’ve heard a lot about currently. I expect development in this area to explode in the near future.