Tartan Data Science Cup

Episode I: Ride To Glory

The Problem

You are a data scientist working with Citi Bike NYC. You just received notice that there is additional funding to build two additional bike stations in New York City.

Your boss wants to position the new bike stations in a way that will increase ridership. He wants to make bike access easier for both subscribing members (so that they continue to be returning customers) and one-off riders (in hopes of convincing them to subscribe and use the system more frequently). Additionally, your boss is particularly concerned that female ridership is low, and wants to position the new bike stations in locations that may increase the number of female riders.

Given the dataset provided below and any external information about New York City, where do you recommend placing the two new bike stations, and why?

Data

The dataset you will work with comes from the Citi Bike NYC system data. Specifically, you only have access to a subset of the full dataset from July 8th, 2015. As such, your presentations will be judged based on your use of this portion of the Citi Bike NYC dataset only. It will not help your cause to use any other Citi Bike NYC datasets for the purposes of this competition.

The dataset contains 15 variables and 35,047 observations, and can be accessed here. The variables are described below:

• tripduration: The duration of the trip, from start station to end station (seconds)
• starttime: The date and start time of the trip (format: m/d/yyyy hh:mm:ss)
• stoptime: The date and end time of the trip (format: m/d/yyyy hh:mm:ss)
• start.station.id: An ID number corresponding to the station from which the bike was taken
• start.station.name: The name of the start station
• start.station.latitude: The (geographic) latitude of the start station
• start.station.longitude: The (geographic) longitude of the start station
• end.station.id: An ID number corresponding to the station to which the bike was returned
• end.station.name: The name of the end station
• end.station.latitude: The (geographic) latitude of the end station
• end.station.longitude: The (geographic) longitude of the end station
• bikeid: An ID number corresponding to each unique bike
• usertype: An indicator of whether the user is a subscriber or a one-off rider (Subscriber vs. Customer)
• birth.year: The user's birth year (available only for Subscriber users)
• gender: The user's gender (available only for Subscriber users); 0 = Unknown/Other, 1 = Male, 2 = Female

Rules

First and foremost, you must use the data provided as the primary source when justifying your answer.

You are welcome to use supplementary information (e.g. maps, info about important locations in NYC, etc) to justify your answer, but you should NOT use any additional data from Citi Bike NYC.

Exactly how you justify your answer is up to you. That said, we suggest the following:

• Use graphics / data visualizations
• Consider the location and usage of nearby bike stations
• Consider the location and traffic of nearby intersections
• Consider the location of nearby landmarks, businesses, tourist attractions, etc
• Consider the location of existing bike lanes in NYC
• Provide a justifiable prediction of how much bike usage would occur at a new station (and any effect this would have on usage at nearby stations)
• When appropriate, incorporate the results of statistical models/tests

Versions of this dataset have been analyzed before in other competitions. You are welcome to look into these for ideas and other information. That said, your final submission must be your own. Any teams suspected of copying/plagiarism will be disqualified at the discretion of the organizers and judges.