Motivation and Inspiration
So I came to Illinois in Fall 2014. Being pretty much new to the city, I used Yelp a lot for exploring places in and around campus. From getting a hair cut to finding the best pizza place, Yelp has been a good source to learn more about a business. But one thing I observed is that I had to read a few reviews before I could make up my mind about the place.
This inspired me to build something that can help the user get an overall idea about the "business" without having to go through many many reviews.
In order to do that I got hold of the Yelp dataset and basically did keyword extraction, sentiment analysis and word cloud generation that will help the user get a quick overview of the place. Thus it will help the users better understand (and hopefully help them!), when they make a search.
- Since I have used the Yelp dataset, for now the system targets any user searching for businessess.
- But I have made it generic such that by plugging in a different dataset the system can be adapted to a different domain as well.
- Addressed and solved real world problem.
- Used elasticsearch to ensure that the system is scalable for future development and can handle increase in data.
- The output to the user is a word cloud that summarizes the key information from the reviews, providing a better visual cue to the user!
- Not giving up when my teammates couldn't make it to the hackathon!
(The information is color encoded to distinguish between Positive, Negative, Neutral and Mixed sentiments expressed in the reviews)