Taming the Twitter

One of the evolving frontiers within business intelligence is gathering insight from the world of unstructured data.  Particularly in the web 2.0 world, data volumes are not just growing but exploding.  Twitter is a great example.  For companies looking for business insight, there’s gold to be found in the world of unstructured data from blogs, Twitter, Facebook, and the list goes on.

Recently we participated in an interesting experiment to build a window of insight within live Twitter feeds.  Using an innovative data extraction technology from Kapow Technologies (www.kapowtech.com), one is able to retrieve thousands of Twitter feeds on any particular word or term.  These feeds are then fed through a text analytics engine (in this case it was Clarabridge) that has the ability to give a quantitative score to each text string (in this case each tweet).  The score ranges between +4 to -4, reflecting whether it was a favorable or an unfavorable tweet.  Empowered with this information, we were able to create powerful visual insight into live streaming Tweets.

What made things even more interesting is that Kapow’s extraction engine could also capture the user location of each tweet, and therefore we could add a layer of mapping intelligence.  See the dashboard below – it shows the Tweet sentiments for Toy Story 3 and Karate Kid.  We used a scaling factor of 33 for sentiment scores to plot on a relative scale of -133 to +133.


(Note: Due to confidentiality of information, we have scrambled the data displayed in this dashboard).

It is a ‘wow’ moment to see how three innovative technologies (information robot, text analytics and data visualization) interfaced to easily create such a powerful solution within a few hours that provides an amazing insight into the world of unstructured data such as Twitter.

Shadan Malik
President & CEO, iDashboards