HackAI2018 -TeamExtremeLoss - Yun Ho Jung , Goostavo Araiza

Using the Kaggle dataset that included 10,000+ information about apps in Google Play Store, we modified the data in a way and trained a K-Nearest-Neighbors algorithm in hopes to make a way of predicting the success of an app.

In the end, we analyzed the impacts of category, price, age rating, reviews, and size of the app to how many installs it acquired, and we were able to predict it with 48% accuracy.

One thing we learned is that KNN may have not been the best way to approach this problem, and we are planning to further attempt to solve this problem using other algorithms as well.

Built With

  • jupyter-notebook
Share this project: