TweetSentiments.com provides Sentiment Analysis on tweets using Natural Language Processing and Machine Learning technologies.
Sentiment Index [100*((positive-negative)/total/2+0.5)]:
most negative |
0 | |
negative     |
1..49 | |
neutral |
50 | |
positive     |
51..99 | |
most positive |
100 |
The profile is estimated based on the recent tweet content and writing style, the Education level could be interpreted as person with that level of education is capable of writing the tweets, and does not necessarily reflect reality.
The sentiment/profile results are calculated using OpenAmplify's Natural Language Processing(NLP) application and LibSVM/LibLINEAR(Support Vector Machines) machine learning tools. Should you have any questions regarding the results, please post them on OpenAmplify's forum, or contact us using the Feedback tab. We continuously improve to our algorithms and the results will get better over time as we process more and more data.
The site is built with Ruby
on Rails(web application framework), Twitter
API, OpenAmplify, LibSVM/LibLINEAR, GIS, and visualization technologies. A lot more features to come.
The internationalization/localization is done automatically using Google Translate, as expected, there are a lot of out-of-context translations, those can be and will be manually corrected in the future(can be done online on the live production site). All Google Translate supported langauges are enabled here.
Please contact us using Feedback(tab on the left), or at tweetsentiments [at] gmail.com for comments, suggestions, bug reports, and feature requests. And follow us on Twitter @tweetsentiments for hourly Sentiment Index(published every 6 hours) and daily Top Tweeting Country list and sentiments updates.
(c) 2009 Tom Z Zeng
http://www.linkedin.com/in/tomzeng
http://www.tomzconsulting.com