Having built TwURLed News to the point where it runs reasonably reliably and does something somewhat interesting (damned by faint praise, eh?), I decide that if there’s any real future in this sort of thing, it needs to reach beyond Twitter. It also needs to be segmented, I suspect, but that’s food for future thought.
The cool thing about Twitter is how open it is, providing plenty of data for those of us who imagine we can discover value in the ways that people are interacting. The less cool thing about Twitter is that although it is popular, it is far from the only fish in the sea. Personally, I get a lot of use out of Facebook, LinkedIn and sometimes Plaxo. And I’m fairly certain that MySpace is doing okay, too. Let’s not forget Digg, either. And, I have realized for the umpteenth time, the web itself. That last source of social network data – the web – came back to my attention as I started taking a look at a long list of social networking APIs on Programmable Web. Wow. I had no idea. Some are definitely more interesting than others.
The one that really caught my attention was Google’s Social Graph API, which returns data about publicly identified connections among people on the Web. It seems particularly useful when querying sites that people use for aggregating their web presence, such as FriendFeed. For example, query my FriendFeed page and you’ll get links to this blog and other places I write and publish. A query on my Facebook page doesn’t yield anything interesting other than the link to my FriendFeed page, which shows the difference between a quite closed system and an open one.
These APIs are itneresting because I’m always curious to see if I can quickly create and maintain smart data exploration robots, who can find useful and interesting information without trying to boil the ocean. Twitter should be just one source in a generalized framework that leverages other peoples’ code and data as much as possible. In fact, I think I’ll probably slow down directly querying the Twitter API in favor of relying on several of the URL tracking sites, such as Tweetmeme and Twitturly. My main concern about that approach is that I don’t know how to rate the risk that such services will continue to exist… and accessing them via their API doesn’t seem to do anything to help them stay in business. That’s a bit of a conundrum, but the only answer I’ve come up with is to avoid reliance on any one of them, just as I don’t think it would be a good idea to focus entirely on just one social network.
As I take a look at the various APIs, I’m trying to block out a software framework that will take advantage of these and future data sources to help me spot interesting themes, trends and so forth. That’s been my goal all along… without boiling the ocean in the process.
Twitter: Web Analytics (#wa)