Sunday, January 14, 2007

Use of tag clouds and words to tell a story

1 comment :
Tag clouds are an interesting way to see trends across large collections of user tagged content like flickr and i am starting to see them being used in the enterprise as a tool to find interesting content, subject matter experts and 'brewing' topics. Enterprises that are using tools like blogs and wikis can easily leverage those technologies to drive tag cloud views. Most tag clouds are generated by the tags that users apply to the content element- be it a picture, video, blog post or any other type of document internal or external to the enterprise. There however are other ways to create 'tag clouds' such as using text mining technologies to extract the most frequently used terms and phrases of specific data sets, or looking at search terms that people are using like what The New York Times Search Tag Cloud does.

Via Richard MacManus's Read/WriteWeb blog comes this
interesting use of Tag Clouds over at Seattle Post-Intelligencer's online site as part of the media coverage they provide online. As newspaper audience continues to shift to online mediums, it is tools like this that brings value add to readers and potentially drives readership. Todd Bishop's blog covers Microsoft and he has posted some interesting views of Bill Gates CES speech. To learn more about this project, read Todd's accompanying story and related blog post as well as some of the comments on his post. There is also a timeline of Microsoft's evolution in words that Bishop and his team produced that is another unique use of tags by adding a slider view feature. The tagline generator is available for use under a Creative Commons license over at

UPDATE (Jan 15th)- Bishop delivers on a reader's suggestion and does a
Bill Gates and Steve Jobs: Keynote text analysis and provide an analysis on Michael Dell as well.

It is always interesting to see journalists like Todd Bishop using words to analyze and understand what is being said about a company, industry etc. especially in such an interactive way for the reader. Factiva iWorks/Search 2.0 has visualization capabilities that utilize text mining technologies to identify top trends, relationships and patterns within your search results and it has become a very popular feature for various types of users from journalists to PR folks because it not only saves time, it discovers topics that might have been missed if the user had to read through all the results. Journalists have also used our text mining platform to do word analysis on specific topics-for example on the clich├ęs that are most used by the media.

1 comment :

Ted said...

You might also be interested in the innovation that we are playing with here at the Personal Bee of creating "Phrase Clouds" -- these are built from phrases extracted from the articles themselves rather than relying on author or audience generated tags. The algorithm focuses on extracting time relevant phrases (we examine phrases against a history for the given source) so that we can identify emergent news trends.

Here is a Web 2.0 example:

The Web 2.0 Workgroup also has a beehive: