Project Name

Predict the demand of bikes for a rental bike sharing company.

Table of Contents

• Steps followed for prediction task.
• Observations
• Conclusions

Steps followed

• Reading and Understanding the data
• Visualising the data
• Splitting the data into training and testing sets.
• Rescaling of features
• Running Recursive Feature Elimination wto get 15 variables.
• Building model using statsmodel for detailed analysis
• Dropping features further on the basis of VIF and p value.
• Linear regression assumptions check
• Model Evaluation

Observations

• 15 features taken from RFE
• Categorical variables seem to have more impact on dependent variable.
• In numeric variables temp features has more impact on target variable though its VIF was pretty high initially

Conslusion

Demand of bikes depends on these variables - yr , holiday, temp, windspeed, mnth_sep, weathersit_Light_snowrain, weathersit_Misty, spring, season_summer and season_winter.