Abstract
Citi Bike is one of the preferred mode of transportation in New York City. It offers timely pass for single trips and daily unlimited trips, as well as yearly and monthly subscription for its users. This study analyze the difference in the trip duration between the Citi Bike subscriber and pass riders. Statistical T-Test is employed to analyze the distribution of trip duration between the two user groups using a trip history data from June 2017, in which the result shows a significant difference between subscriber and pass riders.
Introduction
Citi Bike is a paid bike sharing program that distributes shared bikes across the main neighborhoods of New York City and specifically designed for short-term trips. In general, Citi Bike allows two types of plan with different pricing to access their shared bikes. The timely pass can be purchased as single trip for $3 each or as a day pass with unlimited 30-minute rides in 24 hours for $12. The annual membership with unlimited 45-minute rides can be purchased for $169 yearly or $14.95 monthly.
The different plans cost to access Citi Bike suggest a method of targeting the use of Citi Bike as mode of transportation to two different yet specific audiences. A membership would appeal more to NYC residences as their mode to transport in their daily routine, while a pass would make more sense to short term visitor to explore the city. It is then interesting to see the difference in the use of Citi Bike between membership and non-membership users.
Data
Citi Bike opens its trip history data to the public on its website as monthly dataset from 2013 to 2018. Trip history data from June 2017 is chosen to be used in this analysis in order to give the a better sense of present Citi Bike use where the weather condition is optimal for bike rides in the city. The main variables within this data that is useful for the analysis is user type, which contains subscriber to describe rides by customers with membership and customer for rides by customers with passes, and trip duration for each rides.
The data processing step of this analysis includes separating the data based on user type and filtering the data based on the trip duration limit for each plan (45 minutes for membership and 30 minutes for passes) in order to remove outliers that may come from trip recording error while also pose the assumptions that these riders in the analysis followed the limitation of their plan.