Twittermeter Graphs Twitter Chatter

This weekend while attending Widget Dev Camp in Washington, D.C. I had the opportunity to check out a pretty slick new Twitter application that charts the number of occurences of any given word on Twitter over a period of time. Spikes for any given word are typically a result of some event occuring on or around a specific date. For example if you chart “apple” you will see a spike around the time of Macworld.

Other charts yield interesting yet not surprising results. If you graph Obama versus Hillary, you will see Obama winning by at least a 2-to-1 margin. You can also use the charting for viewing your popularity versus your friends. I found that I have surprisingly active Twitter conversations. I spoke with the developer of this application and he said he is indexing absolutely all tweets on Twitter going back to November.

He also leveraged and open-source MIT project called “Similie” that provides simple graphing services for developers. I was impressed with the speed of queries considering the number of tweets being stored in his database. Twitter’s slow load time may not be due to their database though but instead due to the overhead required to handle the constant text messaging going through the system.

If you’d like to chart yourself against your friends or try to find random anomalies, go check out Twittermeter.