Comments on: Searching Twitter Better

By: Unit Structures – BackTweets

Unit Structures – BackTweets — Sun, 08 Mar 2009 05:55:02 +0000

[...] Something I asked for a long time ago. Don’t know why Twitter search still doesn’t do this, perhaps now they will. Great execution, smart defaults, instantly indispensable for anyone monitoring Twitter. Excellent. [...]

By: factoryjoe

factoryjoe — Sat, 07 Jun 2008 21:32:00 +0000

Seen TwitLinks (http://www.twitlinks.com/)? Seems like it’s doing something like what you’re talking about…

By: fred

fred — Fri, 06 Jun 2008 16:51:00 +0000

@bertil -

I think that’s sort of what I was thinking. The problem with it is exhaustiveness. If you want to do a post-hoc search you need to hit 150+ url services and 150+ searches on tweetscan. So much easier for them to build a fulltext index. TinyUrl, which is the most common in the corpus, has an api…it could be batched. The rest problably send over 300′s so you could build the remainder of the corpus with HEAD’s.

By: Bertil Hatt

Bertil Hatt — Fri, 06 Jun 2008 16:14:00 +0000

Not sure this matches your needs, but this Yahoo Pipe sounds like it has at least the bricks you need:

“Given a Tiny URL, find the original link URL, and then look up on Google web search and Google blogsearch what sites link to the URL.”

http://pipes.yahoo.com/pipes/pipe.info?_id=caa47cb2b123ca36cb4f21256a4eb033

What I assume you can do, is use a hack of this one, in a embedded loop (Y! demands you have two ‘pipes’ whenever you want to loop anything.)

By: Charlie

Charlie — Fri, 06 Jun 2008 14:45:00 +0000

I’ve thought about this, too. Someone needs to dive into all those TINYURLs, URLTEAs and such and pull out the list of where they’re linking to.

Seems like it would be easy to do, too… I’d be curious where all the twit links go.