Indian political Tweets aggregation and analysis

• Main Author: Ng, Bao Zhang
• Other Authors: Anwitaman Datta
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/59173
• Institution: Nanyang Technological University

With the increase usage of using Twitter over the years, more people are using this portal to voice out their opinions and suggestions to their friends. This platform has been utilised by the politicians who are taking part in the upcoming General Election, 16th Lok Sabha in India to interact and share information in Twitter with the general public. As Twitter plays an increasing important role in the political scene thus it should be taken into consideration as a valid source of data analysis during election period. The objective of this project is to aggregate tweets related to the politics in India by capturing related keywords and creating a repository for storing the tweets and for further analysis. Twitter‘s Streaming API is used for this project to retrieve ongoing tweets related to the keywords defined. It also allows an accurate data set for analysis as the tweets being extracted are the most recent data. A total number of 13 million tweets were collected during the period from 10th March 2014 to 17th March 2014. The data collected was based on the general terms used during the Election and name of the political parties. Tweets were extracted with the help of the Twitter search engine and are further classified based on the respective keywords and grouped together to form the base of the project findings. The data collated in this project have predicted that Indian National Congress is more likely to win more seats in this General Election among the 6 national parties due to the fact of having higher frequency count mentioned in the tweets collected. In conclusion, this project provides us with a clearer picture on how Twitter correlates with the political scene. However, further experiment can be done by using different analytical methods to provide for a more comprehensive prediction.