Getcher Visual Tag Display with TagLines

Several months ago I mentioned Zuggest, which I described as Google Suggest, only for Amazon (see that writeup here.) Now the creator of that site, Francis Shanahan, has created TagLines, which you can check out at http://www.francisshanahan.com/taglines/default.aspx?cat=All .

How does TagLines work? Francis describes it this way:

"1) RSS feeds are stored in a SQL Server database.
2) Every few hours, I query the RSS feeds for new articles. These are then stored in my database. Duplicates are weeded out.
3) I take each new article and submit it to the Yahoo Term Extraction API. This is done through REST and Http.
4) I then store the resulting terms in my database. These are the tags."

There are four categories of tags available: Technology, News, Popular Culture, and All. I picked Technology. The tags there ranged from Add to XML and varied in size and color depending on how "busy" and prevalent they were. (Don't quite know what the numbers in brackets are for-- Francis?)

Click on a tag. You'll get a floating window with options to see Shanahan's tags (which I think are the RSS feeds/stories from which he pulls these tags), images, movies, news, and the Web. As you click on each item, a new floating window appears below the existing one. (If you're clicking on tags at the bottom of the screen, you'll need to scroll down.)

Nice presentation and nice collection of tags, though I had a couple of concerns. Some of the image search results came uncomfortably close to being NSFW, so you may want to be careful of that. And unfortunately on the news results sometimes the same story is repeated over. And over. And over. again.

Other than that, it's good stuff. Now what would be the potential of extracting the most relevant terms from Yahoo News RSS feeds -- general ones -- and then checking/mapping them against new/recent/active del.icio.us tags? In other words, would it be possible to discern that there's a pattern for tags in news stories that parallels/follows/preceeds tags in del.icio.us? And if that were possible, would it be possible to discover if there were del.icio.us users who were better/best at preceeding tag use in the news?

Thinking aloud....

Posted to Internet-Technology-Tagging | TrackBack