This project seeks to uncover factors that influence Google Play application installs, which serves as a proxy for app success. Two datasets are merged and analyzed to address two hypotheses:
- review sentiment impacts the number of app installs
- app category impacts the number of app installs
Google play store apps are typically accompanied by user reviews sorted from most to least helpful. Prospective users make the decision to install apps for a variety of reasons, and in this study we seek to uncover the relationship between review sentiment and installation rate. The following questions is addressed: “How much influence does review sentiment have on overall installs?”
Two datasets were obtained from Kaggle and Datacamp.
- apps.csv: contains all the details of the apps on Google Play. These are the features that describe an app.
- user_reviews.csv: contains 100 reviews for each app, most helpful first. The text in each review has been pre-processed, passed through a sentiment analyzer engine and tagged with its sentiment score
Exploratory, ANOVA, and Regression analyses are conducted to adress the project question.
- Games , Productivity and News and Magazine apps have significantly higher Install rates
- 0.1 increase in sentiment score leads to 5.75M decrease in installs (p=0.002)
- Increase of sentiment score controlling for app size, price, and category is not significant
- App sentiment is not a reliable metric of its success - further analyses using in-app data use may be a better metric
- Negative reviews should be used to support product improvement rather than be used as a metric for it's success.
- Higher utility apps will have high installation rates (including as news & productivity apps)