Applytics

Sentiment Analysis of Your App

We developed this project in a hackathon to perform sentiment analysis on reviews of mobile apps on Google Playstore. The key idea was to identify negative reviews quicker so as to avoid user churn.

Training Data:

We scraped reviews of 23 popular mobile apps and manually labelled each data as positive or negative to create the dataset.

Model

The project used Naive Bayes Classifier to perform the sentiment analysis which gave 85% accuracy on our validation set.

Venue

Developed at HackClash-Nepal 2016 hackathon on 2nd & 3rd January, 2016.