Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Fetching latest commit...
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

Twitter Sentiment

Take a tweet, extract a TON of information out of a short bit of text. Let's do this.

Start from scratch

Install RVM (if you haven't already)

Follow the instructions at In short, install RVM:

$ bash -s stable < <(curl -s
$ source ~/.bash_profile

Linux Users

$ rvm install 1.9.3 

Mac Users

Install XCode from the App Store if you haven't already (it's free).

$ rvm install 1.9.3 --with-gcc=clang

OS X users may be interested in Jewelry Box, a Cocoa UI for RVM.

Install the necessary gems

(don't worry, # is a legitimate bash/zsh comment so you can still copy-paste)
gem install rake            # Ruby Make - build tool
gem install yard            # YARDoc docuementation generator
gem install yajl-ruby       # Fastest JSON parser this side of the atlantic
gem install cucumber        # Cucumber BDD/TDD test suite
gem install tweetstream     # Ruby wrapper for Twitter Streaming API
gem install twitter         # Ruby wrapper for Twitter RESTful API
gem install face            # Ruby wrapper for Face (recognition) API
gem install paint           # Ruby pretty colorful console output
gem install progressbar     # Ruby pretty console progress bars

We will move this to a RubyGem when development is further down the road to make these dependencies more easily fulfilled.

Directory Structure

dict/ contains the collection of dictionaries (bag of words, or BoW) being used for sentiment analysis.

lib/ contains the generalized libraries used by our toolkit.

research/ is a general placeholder for interesting papers and potential BoWs.

Research Leads



AFINN: A new word list for sentiment analysis

Simplest Sentiment Analysis in Python


Twitter Mood

ANEW Sentiment-Weighted Word Bank

Measuring User Influence in Twitter


Sentiment strength detection in short informal text

Twitter as a Corpus for Sentiment Analysis and Opinion Mining

We Feel Fine

Modeling Statistical Properties of Written Text (lookup!)

Something went wrong with that request. Please try again.