Twitter's Trending Topics only go as far as showing what people are talking about. To find out what's happening, you have to read through all the tweets about that trend and try to figure out whether people are happy or sad about that topic. In addition to that, many of the top trends are about the same topic but are still separated because they are different keywords. I want to solve this problem. I want people to be able to just think about one topic and they will automatically be shown what that keyword and other related keywords are, and tell you what people are thinking about that right now.
What it does
Language Modeling is used to analyze what people are thinking about on certain topics. It analyzes what people are tweeting about on that topic and does a real-time predictive analysis.
How I built it
The model takes a keyword and then uses Haven on Demand's Find Related Concepts API to identify the entities used in the Twitter search. It then streams Tweets to find all the related concepts to the entity and does sentiment analysis using HavenOnDemand Sentiment Analysis API and finds the aggregate sentiment of the tweet on a scale to 0-1.
Challenges I ran into
Pipeline of the application took some time.
Accomplishments that I'm proud of
Learned how to use HavenOnDemand APIs.
What I learned
Deploying a web application using Node.js.
What's next for Language Modeling
Storing all the tweets in a database and doing predictive analysis on the average sentiment of the tweets on the topic using HavenOnDemand Predictive Analysis API.