Abstract
I have tried to make conclusions about consumption patterns of a section of the bicycle subscribers' base in the city of New York. A finding of such a mini project as this, could, for example, help authorities plan for increasing the number of durable bicycles to be available in every single station, for a slice of the bicycle riding population. This observation –pivoted on the assumption that the riding styles of adults, here filtered between ages of 20 and 40, could result in higher attrition to bicycles – tries to make a suggestion for providing certain user groups with durable bicycles; for minimizing maintenance costs in the long run and perhaps helping the citibike share system become more efficient. My motivation for doing this project was that, I thought people in my sample could benefit from varied types of bicycles, for their daily use. I wanted to understand whether or not bicycle types have any bearing on the overall riding experience of users in my analysis.Unfortunately the outcome of my project confirmed that my hypothesis can be rejected.
Introduction
Citibike is a subscription based shared bicycle transportation system in NYC. Subscribers could find bicycles docked at several stations around the city. My problem formulation tries to examine trip patterns that center around the age group of 20 to 40 years old as I wanted to know if people in this age group were more likely than others to take longer trips of duration, which is more than half an hour. Based on my results the hypothesis that the age group in question rode more than half an hour, more frequently, was rejected. To casual observers it feels differentiating among citibike system users and making bicycle types more varied in terms of their levels of durability and even speed, could be beneficial to a certain section of the ridership.
Data
I have used citibike trip data for January 2016. In that month a total of 509, 478 trips took place around the city. I found the total number of citibike subscribers between ages of 20 and 40 to be 235539, and that of other age groups to be slightly higher, 273939. As far as data processing goes, I used basic data wrangling techniques; I needed to extract the ages of bicycle riders in January of 2016 together with their trip duration.