Predicting Bike Rentals with Happenstance Data

Another interesting and challenging project. Three different machine learning models were used. Prediction errors were compared. When analyzing happenstance data for predictability, typically error would be high. Find out why.
All feedback is appreciated. Thank you kindly.

Predicting Bike Rentals.ipynb (861.9 KB)

Click here to view the jupyter notebook file in a new tab


good job!

random forest does better because it captures better non-linear relationship are the kind of relationships that happen most in real life. Try to visualize through a scatter plot the dependent variable with the independent ones. Try to visualize the residuals as well and scatter them with independent variables as well to see if any one of them is highly correlated to the errors. If not, the omitted variables bias might mess your model which you took into account and there is not much you can do about it.

I would 46% R-squared is not too bad, but the error rate is concerning. Try to standardize your features.

Good luck!

1 Like

Thank you Luca for your input and suggestions! Much appreciated.
Regards, Bruce

This is super cool! I hope to get to this project in the course at some point, I just started so it is nice to see what is possible using Python.

1 Like