This assignment is a programming assignment to build a multiple linear regression model for the prediction of demand for shared bikes.
- EDA
- data preparation
- import linear regression libraries
- train test split
- scaling
- correlation heatmap
- modelling
- Inferences
- Temperature seems to have a good positive correlation with bike sharing. It’s possible that the USA being a cold country, people would like to use more bike sharing when the temperature goes higher
- Rise in snowfall or rain seems to have a moderately negative correlation with bike sharing. It makes sense that when it snows or rains, people wouldn’t want to ride a bike
- Year by year bike sharing is increasing. It's possible that after the pandemic, people are moving out more and using bikes.
- Since cnt was scaled, we should use
scaler.inverse_transform('cnt')to get the unscaled value of cnt - m12 is our final model.