What is the best regression model you can find for predicting total number of bikes rented in a particular hour (variable cnt) using environmental conditions and other available explanatory variables?
What is the interpretation of your chosen model, and how much do you trust that interpretation? Is it reasonable to view any of your fitted parameters as causal effects? Why or why not? What guidance does the form of your model provide for future studies of bike share use?
Some of you may be tempted to use sophisticated machine learning methods or time series approaches (since the data has a time series structure) for your analysis. Please remember that this course is about linear regression.
The dataset contains information on bike rentals for two years (2011 and 2012) from Capital Bikeshare System, Washington D.C. Bike sharing systems are bike rentals where the whole process from membership, rental and return is automatic. Through these systems, the user is able to easily rent a bike from a particular position and return at another position.
This dataset is collected to address the problem of predicting the number of bike rentals in a given hour given the environmental and seasonal conditions for that hour. The dataset contains 17379 observations with each observation corresponding to one particular hour. The dataset contains the following 17 variables:
- instant : Unique observation number.
- dteday: Date.
- season: Categorical variable (1: Spring, 2: Summer, 3: Fall, 4: Winter).
- yr: Stands for year. Binary variable (0 stands for 2011 and 1 stands for 2012).
- mnth: Stands for month. Takes the values 1, 2, . . . , 12.
- hr: Indicates the hour of the day (takes values 0, . . . , 23).
- holiday: Indicates whether the day is a holiday or not
- weekday: Day of the week, numbered 0 (Sunday) through 6 (Saturday)
9.workingday: Takes the value 1 if the day is neither weekend nor holiday and takes the value 0 otherwise.
- weathersit: Takes four values:
(a) 1 if the weather is Clear, Few clouds, Partly cloudy, Partly cloudy.
(b) 2 if the weather is Mist + Cloudy, Mist + Broken clouds, Mist + Few clouds, Mist.
(c) 3 if the weather is Light Snow, Light Rain + Thunderstorm + Scattered clouds, Light Rain + Scattered clouds.
(d) 4 if the weather is Heavy Rain + Ice Pallets + Thunderstorm + Mist, Snow + Fog.
Sample Solution