During the process of making the maps for the last assignment in Urban Informatics and for this post, I created a map, which looked like this. My immediate thought was, this does not show bike trips but the smell of Manhattan, caused by the trash, industries and many cars inside Manhattan. However, this is not the case; this dark mess are the routes citi-bike users are taking.
Although the air pollution in New York City is definitely worth its own thorough research, I will focus today solely on citi bike data, again. This time, however, I will look at the top ten start location, which citi bike users will use. The month of interest will be September 2018.
The red graph here displays the number of trips per hour from the entire month. As expected, the lowest use of bikes is during the night between midnight and 5 am, while the peak points are in the morning (7 am to 10 am) and in the late afternoon/early evening (4 pm to 7 pm), most likely due to the people commuting to and from work respectively. The interesting part is the constant rise from 10 am to 4 pm during workhours. The main factors for this rise might be students or people who starting work late or leave early. For future research it would be interesting to analyse the kind of tripes, for example commute to work or university, if the same person uses citi bike in the evening again or decide to use another mode of transportation.
In the following analysis I look at two time intervals, the morning and the evening. I will not differentiate between weekdays and the weekend. This offers a quick overview of the starting and ending points of each citi bike trip. Even without a deep knowledge of New York’s geography, there are several geographical features visible. Once, Manhattan is the main area in which the nodes are located. In the northeast with no nodes, Central Park is recognisable. North of Central Park are the Bronx. East to Manhattan is the East River, separating Manhattan from Queens and Brooklyn. The other area in Brooklyn with only few nodes is home to the Newtoon Creek.
With those geographical features the figure displays several musters. There is little to no interaction between Bronx and Queens, only between Bronx and Manhattan. To be more precise, the south of the Bronx. The north Bronx do not show any citi bike data. A similar pattern is present for Queens as well, albeit Queens has several connections to Brooklyn. The two major players of this dataset are the cyclists from Manhattan and Brooklyn. An overwhelming number of nodes and trips are visible in Brooklyn and in Manhattan and between both of them.
The table shows the nodes with most connections. Those nodes are both starting and ending points. Every single location is in south or in southwestern Manhattan, except for Kent Avenue & North 7 Street, which is located in Brooklyn at the East River. Given the characteristics of those neighbourhoods, the assumption of my previous blog post, that “especially wealthier NTAs will use bikes more often” than other NTAs seems to be confirmed by this study as well. The eigenvector centrality, displaying the influence of the nodes on others, is showing the same results.
Looking now at the evening data it will be interesting to see if the visualization of the nodes and trips differ from the morning and which node will be on top. If there is no difference, it would be more logical to assume, that the majority of the citi bike trips are used for short travels and not as a mean for commuting to work or university.
To my huge disappointment there was no difference in the graphics nor in the centrality measures. Although the original data table was split in two data tables containing morning and evening respectively, the results of several analyzation attempts remained the same. While there is the chance of the datasets being identical by chance, I do not assume so, but rather suspect that there is a logic and/or typing error. I will look into the code into more detail in several weeks, after I gained the necessary mental reset for this data.
Nevertheless, the result of this analysis is clear. Only households or NTAs with a higher income are willing or able to take a bike for commuting or in their leisure time.