New York start-up SocialGuide has launched from beta and released its first television ratings report this week, based on information mined and filtered from more than 10.5 million social media comments by more than 2.6 million unique users.  This report, the Social100, gets most of its information from Facebook and Twitter, using application programming interface (“API”) streams to capture real-time social media comments on 4,150 television shows.

According to SocialGuide:

Our proprietary Intelligent Social TV Recognition System uses programmatic rulesets to dynamically create keywords and phrases about a specific program that we use to identify potential social conversation about a TV show. We then use additional natural language processing techniques to identify the “Social TV Comments” and “Social TV Uniques” of programs, matching them to specific episodes or program events – as they air within their timezones. Our editorial staff further augments our efforts by manually reviewing thousands of the most popular TV shows.

SocialGuide is far from the only start-up operating on a business model that relies on gathering information from API streams.  Of particular note is GNIP, which launched from beta in 2010.  This API aggregation company combines data from more than 100 social media sources into a single API and sells access to this data to other companies that wish to monitor social media, typically for marketing purposes. 

SocialGuide’s television ratings have begun to garner attention from mainstream press and from the television industry.  However, so far only the tech industry has focused on the issues surrounding the technology underlying SocialGuide’s rating system, namely, the sharing of user information between social media and other companies, using API.