The Awards Meter: Analyzing Big Data in Real-Time

The annual movie awards season is always a source of hours of television programming, media buzz and conversation. Social media platforms like Twitter have now made it possible for millions of people all over the world to join in the prediction dialog. Millions of words of commentary are published before and during the awards telecasts speculating on who and what movies will win.  A team at HP Labs has been conducting research on collecting and measuring different forms of social media and presenting real-time results based on live feeds from sources like Twitter.  During the recent 2013 Oscars telecast, the HP Labs social media analytics team together with a team from the HP Enterprise Services CTO office, hosted an implementation of this research, called Awardsmeter (http://www.awardsmeter.com) on the HP Cloud.  Awardsmeter provides a real-time leaderboard of social media “buzz” based on analysis of millions of Twitter feeds.

10,000 Tweets Per Minute

At its peak, the Awardsmeter system analyzed almost 10,000 tweets per minute and processed 3  million tweets related to the award categories during the entire event without a single hiccup!  Supporting the HP Labs software was an HP Vertica analytics engine that was used to parse the incoming feeds and produce only the targeted information.

Analytics use cases like this are ideal for cloud deployment due to the time-limited, elastic nature of the underlying system requirements. Cloud resources can be scaled up to handle a tremendous but short-lived demand for data analysis and scaled down after the event is over. The HP Cloud team is working closely with a variety of analytics systems including HP Vertica and Autonomy to offer innovative, high performance, cloud-based analytics services and support. More product information will be available in the near future.