Twitter_stream is a project mainly written in PERL and SHELL, it's free.
Connects to Twitter streaming API and stores URL mentions
This is my idea:
I was planning on expanding the shortened URLs and doing a HTTP HEAD on each url to pull out useful info, like mimetype, date and such - and if it is HTML I would try to download a chunk of the start to get the /html/head/title.
With that I can put the URLs (shortened and full) into a graph structure with the hashtags and hopefully make a discovery website on top of it with rss-feeds for hashtags.
An SVG/Flash/Raphaël graph traversal UI would be cool.
And then incrementing the URLs as more people mention them incrementing a counter.
The data is not just from my timeline - it's the firehose of twitter (if I had full access, now it's just ~ 5% of tweets).
Imagine being able to take a bit.ly URL you get from somewhere, look it up in a service like that, see what hashtags are related to it, then clicking on a hashtag that you find interesting and then finding more interesting links related to that hashtag.
Somewhat similar to pingwire.com, but for links.
I'd love to have the #perl or #modernperl RSS feed based on this somewhere.