Shared bikes: A prediction based on wheater data
It is an intersting dataset in which wheater data was recorded together with the use of shared bikes. The XGboost regression model got 95% r2 score.
Libraries
- numpy,
- pandas,
- matplotlib,
- sklearn,
- scipy,
- tqdm,
- seaborn,
- datetime,
- math,
- warnings
Motivation
My major motivation was my personal use of shared bikes. As a user, I did not understand why sometimes all the bikes are gone at the station close to my home. The project here give a better idea of when the bikes are used.
Files
blog_regression.ipynb - jupyter notebook containing the analysis
hour.csv - csv file containing the data sampled hourly
bikes_day.csv - csv file containing the data sampled daily
Summary
- Data exploration and explotatory analysis
- Showing that temperature affects bike renting daily and working day the hourly data
- Two models that predicted the amount of bikes daily and hourly with great precision