radialflux / twitterdataanalysis Goto Github PK
View Code? Open in Web Editor NEWThis project forked from saumyasoman/twitterdataanalysis
In this project we did sentimental analysis on data collected from the social media, Twitter and predicted the current trend. The data can be tweets, quoted tweets and the favorites for a tweet (the number of times a tweet has been liked). Data was collected for a pair of keywords using the Twitter Search API. The collected tweets are then classified as positive, negative, neutral or junk based on the sentimental analysis of the text in the tweet/quoted tweet (favorites are considered as positive). Based on this classification it is possible to predict which among the pair of keywords is more popular. The prediction is under the assumption that more positive and neutral responses are there for a keyword, more trending it is with the public. An Android app was created to display data analysis results for a pair of keywords The accuracy of prediction was examined by predicting the outcome of November 5th Governor Elections in New Jersey using keywords Barbara Buono and Chris Christie.