Sigmoid The Dons Final Project
Mobile Apps Market is overwhelmed with games. Even more so - overwhelmed with customer reviews for the games. On average 629 apps are released daily on the App Market. From that 14.2% are in the games category on the PlayMarket. For a developer it is difficult to extract useful data from this amount of information. Our research will ease their jobs by categorizing the reviews not only by user satisfaction, but specific indices such as bugs and features, that can then be used to rate the success of the game and predict its longevity.
For more information about the mobile game market please consult our Business Intelligence Research.
This study researches sentiment analysis for Play Marker game apps. The goals of the study are:
- Build a predictive analytics model to identify critical issues such as bugs, features and “irrelevant” data in real time
- Understand game satisfaction trends
- Analyze game reviews and classify the sentiment of the review into positive, negative, neutral
This project analyzes and classifies the sentiment on a dataset from kaggle. We trained and then tested on a dataset scrapped from Google Play Market with the help of Google Play Scraper Library. The labeling of the data into bugs and features was done manually on 5000 items from the scrapped dataset.
- Data cleaning - spell checking and removing punctuation, stop words and hapaxes
- Stemming
- Tokenization
- accuracy_score(y_test, y_pred)
- output: 0.87
- accuracy_score(y_test, y_pred)
- Output: 0.97
- accuracy_score(y_test, y_pred)
- Output: 0.92
- Play Market Scraper
- included in the requirement.txt document
- Install python virtual environment (conda)
- Run in the shell
pip install -r requirements.txtto install all necessary libraries. - Open the setup.py, change the value of X to your desired input and run the script.
- You will get one of the following texts: irrelevant = 0, bug = 1 or feature = 2