This document summarizes a Kaggle competition to forecast bike share system use using historical rental data. It describes the data provided, which includes date/time details, weather factors, and counts of casual, registered and total rentals on an hourly basis. It also lists the tools that were used for the analysis, including R for feature engineering and models like random forests and neural networks.

1. {
Kaggle <- BikeShare
Taposh Dutta Roy
Jan 26th 2015
Presented at YHAT, Oakland, CA

2. Contents
• About Bikeshare
• Data
• Tools
• R
• Factor Engineering
• Matrix
• Random Forest
• Neural Nets

3. About Bike Share
Competition: http://www.kaggle.com/c/bike-
sharing-demand
Challenge:
Forecast use of a city’s bike share system
Data :
UCI Machine Learning Repository

4. Publication :
Fanaee-T, Hadi, and Gama, Joao, Event
labeling combining ensemble detectors
and background knowledge, Progress in
Artificial Intelligence (2013): pp. 1-15,
Springer Berlin Heidelberg.
About Bike Share

5. Data
The goal is to predict counts either
based on sum of casual & registered or
directly

6. Data Fields
datetime - hourly date + timestamp
season - 1 = spring, 2 = summer, 3 = fall, 4 = winter
holiday - whether the day is considered a holiday
workingday - whether the day is neither a weekend nor holiday
weather –
1: Clear, Few clouds, Partly cloudy, Partly cloudy
2: Mist + Cloudy, Mist + Broken clouds, Mist + Few clouds, Mist
3: Light Snow, Light Rain + Thunderstorm + Scattered clouds, Light Rain +
Scattered clouds
4: Heavy Rain + Ice Pallets + Thunderstorm + Mist, Snow + Fog
temp - temperature in Celsius
atemp - "feels like" temperature in Celsius
humidity - relative humidity
windspeed - wind speed
casual - number of non-registered user rentals initiated
registered - number of registered user rentals initiated
count - number of total rental

7. Data
datetime - hourly date + timestamp
season - 1 = spring, 2 = summer, 3 = fall, 4 = winter
holiday - whether the day is considered a holiday
workingday - whether the day is neither a weekend nor holiday
weather –
1: Clear, Few clouds, Partly cloudy, Partly cloudy
2: Mist + Cloudy, Mist + Broken clouds, Mist + Few clouds, Mist
3: Light Snow, Light Rain + Thunderstorm + Scattered clouds, Light Rain +
Scattered clouds
4: Heavy Rain + Ice Pallets + Thunderstorm + Mist, Snow + Fog
temp - temperature in Celsius
atemp - "feels like" temperature in Celsius
humidity - relative humidity
windspeed - wind speed
casual - number of non-registered user rentals initiated
registered - number of registered user rentals initiated
count - number of total rental

8. Data
Datetime - hourly date + timestamp
Predefined Factors:
season - 1 = spring, 2 = summer, 3 = fall, 4 = winter
holiday - whether the day is considered a holiday
workingday - whether the day is neither a weekend nor holiday
weather –
1: Clear, Few clouds, Partly cloudy, Partly cloudy
2: Mist + Cloudy, Mist + Broken clouds, Mist + Few clouds, Mist
3: Light Snow, Light Rain + Thunderstorm + Scattered clouds, Light Rain +
Scattered clouds
4: Heavy Rain + Ice Pallets + Thunderstorm + Mist, Snow + Fog

9. Data - Continuous

10. Data
Workday busy hours

11. Data

12. Data

13. Tools
Weka
R
Python
H2O + R
Vowpal Wabbit

14. Using R

15. Feature Engineering

16. Citations
Feature-Weighted Linear Stacking
Joseph Sill1, Gabor Takacs2, Lester Mackey3, and David Lin4
Combining Predictions for Accurate Recommender Systems
Michael Jahrer ,Andreas Töscher ,Robert Legenstein