interested in combining your passion for politics with your interest in research, network cluster visualization and unstructured data and text mining? me too! …i was turned on to this sweet site recently by a (now) fellow insna member…it’s hosted by crawdad, a consulting firm currently developing software for commercial release…the site is called the ‘campaign tracking dashboard’ and it is intense once you spend a few minutes reading through the complete explanation of how it works and displays what’s on your screen (yes, you’ll need a computer for this)…

the short explanation from their site: “For the Bush and Kerry sites, we collect and analyze two sources, the news releases (News) and web log (Blog) posts. We locate the releases by scanning web pages, and posts by monitoring RSS feeds…We check the feeds several times daily. When there are new releases/posts, we analyze them and add the results to a database that drives the daily report. We produce a new report each morning…

…Analysis is based on Centering Resonance Analysis (CRA) a patent-pending network text analysis technology developed by Crawdad. CRA performs linguistic analysis on text to identify the words representing the main subjects and objects of discourse. It then links these words into a network that reflects the intent of the writer. It then determines the influence of words by their structural positions in these networks. All of the measures in the report (except for the word/story counts) are based on the CRA networks and/or the word influence values derived from them.”

…from there, the dashboard displays several tracking items over the week primarily around ‘focus, tone, and intensity’ (from a database of over 32 thousand articles)…this makes it possible for the system to compare site text to what would be considered ‘normal text’ …but is there such a thing as normal text in political discourse?

Some similar nonsense, if you like that kind of thing: