Inspiration : Election is the hot topic now. We thought analysing the user opinion based on their tweets.
What it does
Sentiment analysis and depiction of user opinion about a topic based on the tweets by the users. We show a gradient map depicting the temperature of the user opinion. Infrastructure:
How we built it
It is a Stanalone Java application, which users twitter4j to fetch the live tweets by the user on specific topics and specific Hashtags.
- The data is persisted in a Database hosted on an MS Azure VM running a MYSQL server.
- The tweets are analysed for positive or negative opinion about a topic(say elections) using the stanford core NLP API.
- These opinions are rated on a scale for the users at specific locations and shown on the map using D3.js API.
Challenges we ran into
Sentiment analysis tools were not that accurate.