Fishing for insights in the Political Stream
WHAT IS IT ?
Political Streams was a project that was near and dear to my heart during my time at Microsoft Live Labs. The brainchild of the very talented and forward thinking social media scientist, Matthew Hurst. Political Streams provides a view into the ever changing political landscape on a minute by minute basis. The system can crawl a wealth of social media including blogs, message boards, news sites, twits, etc and performs a ton of analysis to pull out relevant entities, tie together conversations, infer sentiment and a whole host of other data mining techniques. I’ve been fascinated with data mining ever since my days back at IBM looking for trends around how people used their Olympics website. Matt had previously worked on Blogpulse, which Nielsen was smart enough to acquire a few years back to broaden out their suite of tools for monitoring and reporting trends and usage in social content. I’d always hoped to see this kind of content available as a module on MSN or Live.com as a more sophisticated feature for newshounds than Google Trends. Maybe with a little more refinement, Social Streams will provide the ability to see the breaking news before anyone knows about it and the tools to get a depth of understanding and context that one wouldn’t receive from any one single source.
WHY IS IT RELEVANT ?
While the world of politics is the first demonstration of the platform, Social Streams can be tailored to analyze any topic such as movies, product launches, celebrities, news, health, etc. Live Labs is staying true to its mission and exploring the possibilities of the research in a public forum. Nielsen Buzzmetrics makes its money by selling reports that are typically purchased by political groups, movie studios and product marketers to gain a better understanding of how consumers are reacting to the introduction of a new product, the ending of a movie or even the presidential debates. The breadth of data that is collected coupled with actionable near real-time insights allows for changes of messaging on the fly, or early warnings on defects or dissatisfaction with a product. Early adopters and influencers can be easily identified by the number of inbound and outbound links to their blog posts. Sentiment can be inferred as to whether people reference an entity either positively (this camera rocks!), or negatively (the movie ending sucked!). The potential for customization and personalization of content and advertising using such as system is massive. Imagine knowing where a great new restaurant is in your area based upon people twittering about their meal 10 minutes ago. Companies like Evernote and Jott are making it easier to collect all your personal information in one place. Performing data-mining analytics and coupling them with the wider web could yield information, connections and insights that you would never have found as an individual. You could be the new trend setter and not even know it…yet.