I did a talk about a Real Time Sentiment Analysis Application at the Hadoop Summit 2012.
Here are the slides from this presentation:
This is an application that evaluates the sentiment of twitter users towards a small number of pre-determined keywords, stores them in HBase, and displays a graph of sentiment versus time. Users can scroll back and forth in time to view how the sentiment tracked over time.
Download the java files and the html files for this project from the links at the top of this post.
There are three parts to this program.
- Using the twitter API to get a stream of tweets (public status updates)
- Doing sentiment analysis on the tweets and storing them in HBase
I am co-founder of AltoStor – we develop software that turns HBase and Hadoop into a metered, billed service in the public cloud or in Enterprise VMware. I used our HBase Workbench to develop this entire project. The code itself will run on stock Hadoop 1.0.x with HBase 0.92.x – you don’t need the Workbench. The Workbench is simply the easiest way for you to get started developing Big Data applications.
Good luck playing with these technologies