[Python] Why is my R² score negative on a test set even though it’s near 1 on training?...

I'm working on a linear regression model to predict the average number of daily commits to a GitHub repository based on Stack Overflow activity metrics (like views, questions tagged, etc.).

My dataset includes features like:

monthly_views (number of times the tag is viewed)

questions_asked (number of new questions tagged)

questions_answered (number of answered questions)

tag_followers (number of users following the tag)

I split the dataset like this:

X_train, X_test, y_train, y_test = train_test_split(X, y, shuffle=False, test_size=0.2)

Continue reading...