Sentiment Analyzer

A Python module to do a set of operations on tweets. It uses a collection of stopwords to train a dataset for the sentiment analysis. It uses the basic principle of bag-of-words used for natural language processing.

Dependencies

• numpy
• matplotlib(To plot sentiments)

How to use

from src import features, datalink, hashtags
import time
dblink = datalink.DatabaseConnectionDown('perilipsi_tweets')
emoTest = features.Emoticons()
dictTest = features.DictionaryTest()
hashtest = hashtags.hashtags()
testTweet, tweetTime = dblink.fetchTweet()['tweet'], dblink.fetchTweet()['time']	#You can pass anything you want
emo_test = emoTest.analyse(testTweet)
dict_test = dictTest.analyse(testTweet)
hash_test = hashtest.analyseHashtagTweet(testTweet)
print "Emoticons:", emo_test
print "DictionaryTest:", dict_test
print "Hashtags: ", hash_test

Output

Emoticons: {'positive': 0.33, 'negative': 0.66}
DictionaryTest: {'positive': 0.46153846153846156, 'negative': 0.5384615384615384}
Hashtags:  {'positive': 0.38, 'negative': 0.62}

Progress

• Emoticons: This class uses emoticons detection to classify the passed string as positive or negative
• DictionaryTest: This class uses a set of English words and their subjectivity to give a score to a string
• hashtags: This class extracts hashtags from the string sent and calculates the sentiment based on a trained dataset
• AllCaps
• ElongatedWords
• Negation
• Punctuation

Social Network APIs

• Twitter Search API
• Facebook Graph API

Computation Engines

Wolfram Alpha

Team Members

Name	Email
Sudhanshu Mishra	[email protected]
Ambar Mehrotra	[email protected]