iliaromanov / google-play-app-rating-prediction

Project With Partner; Using Data Analysis/ Visualization and ML to predict the rating an App would get on the Google Play Store

Hi, I'm Ilia!

• 🎓 Fourth year Computer Science student at the University of Waterloo.
• 👨‍💻 Interested in backend, computer networking, and algorithmic trading.
• 🏢 Looking for Winter/Summer 2025 internship or 2025 new grad Software Engineering related opportunities.

google-play-app-rating-prediction's Issues

Remove Unneeded Columns

First, we want to choose which columns we will keep for our app_data_x variable (the df that we will feed to the model to predict the rating). Probably Category, Size, Type or Price(not sure if we should keep both), Content Rating, Genre, and maybe Android Ver.

Then we should remove the unneeded columns.

Clean Genres

Since the Genres column consists of semicolon-separated arrays/lists (not sure what they're actually called) we should find a way to consider each individual genre in the list.

I am not really sure how or when we should do this.

Maybe create a duplicate row for each genre in the list for that row??

We could drop Genres all together if this is too difficult and just use Category instead

Clean Rating

We need to remove all rows with a null value in the rating column

Clean Size

Convert the values in the Size column to numeric and remove rows with unusable values in this column such as "varies with device" or "nan".

I kinda wanna do this one cus it's one of the only ones I think I have a clear idea for.

Clean Android Ver

If we are going to keep the Android Ver column for our app_data_x variable, we should probably find a way to either convert it to numeric or combine similar values (e.g. combine "4.0.3 and up" and "4.0.3" to be one category). I'm thinking combining similar values would be better because the Android version seems like a more categoric piece of data.

Then have to remove any unusable values from this column such as "nan" and other unusable values of that sort.

Remove Duplicates

Remove apps with duplicate names to make data more readable and clean