Research

The COVID-19 pandemic has disrupted transportation in New York City including: transit, private vehicles, ride-sharing, and bike-sharing. To better understand the effects of COVID-19 on a specific mode of transportation we focused on the bike-sharing system because of its growing popularity in urban cities around the world. During the early stages of COVID-19 bike-sharing in New York City had a peak in ridership. However, with the enforcement of social distancing and the stay-at-home order, the access to Citi Bike bikes had been limited, which has decreased bike usage and mobility around the city. To study this infrastructure, we hope to gather data from the Citi Bike database. We plan to use programming and data management tools, such as Python, SQL, and Pandas to depict bike usage trends prior to and throughout the pandemic. Additionally, we plan on identifying any fluctuations in bicycle “hotspots,” popular destinations, throughout the city. A possible method to locate hotspots is through the Citi Bike database and map oriented routes. We will also make use of extensive literature review to help us understand the main changes in behavior due to the pandemic. We hope this research will help contribute to our understanding of how the bike-sharing system functions and adapts to situations that impede transportation. Ultimately, our results will be used in the ongoing research on the impacts of COVID-19 on New York City’s transportation network performed by the UrbanMITS laboratory in conjunction with the C2SMART Center.

Before COVID-19, biking has caught on as a popular mode of transportation in cities due to environmental and societal factors (DeMaio, 2013). In New York City, the introduction of the Citi Bike bike sharing system attracted tons of people, both customers and subscribed members. The reason why the system was able to attract so many people was due to the convenience of the bikes. It provided transport for quick trips, as well for trips that may be too long on foot (Buehler & Pucher, 2017). Due to this sustainability, Citi Bike stations were implemented throughout the city, most notably in the borough of Manhattan.

The growing popularity of the bike-share network compared to other modes of transportation like the limousine or taxi service. Researchers that compared the two systems created "hotspot" maps which identified the popular places people took bikes and taxis (Ding, Keler, & Krisp, 2019). We incorporated this idea to create our own list of top bike stations prior to and during the pandemic to indicate whether there would be a change in people's popular destinations.

At the early stages of COVID-19, most transportation in New York City plummeted in response to the stay-at-home orders and social distancing. However, research showed that bike sharing systems were more resilient than other forms of transportation like the subway system (Teixeira, Lopes, 2020). Teixeira compared the subway and bike stations through a map-oriented approach. Their research noted that as subway ridership decreased, bike ridership increased. Similarly to the research, we utilized a map-oriented and data-driven approach to understand the stress of COVID-19 on the bike sharing system.

While Citi Bike bikes have been successful, most stations are focused in Manhattan and Brooklyn, and are not widely available in other boroughs (Kaufman & O'Connell, 2017). Since Citi Bikes have proven to be a safer form of travel, implementing stations in other neighborhoods can increase mobility in other parts of the city and expand the outreach of the bike-sharing program.

We used multiple programming languages in order to obtain our results. For our project, we used Python and Pandas Library in order to import and analyze our data. We were able to accomplish thus by using different code commands in order to simplify the data to the information that we need. By accomplishing this, we were able to identify important components such as Top 10 Station ID’s, Daily Ridership from certain time periods, as well as finding out the total number of trips per Station ID.

These components were then programmed into 4 different visuals (two per respective year). The first visual is a line plot that follows the overall trend of ridership from March 1st through June 30th. The second visual is a bar graph that shows the Top 10 Stations for 2019 and 2020 along with bars in descending order displaying the amount of trips per station.

To design an effective and interactive map, we used Mapbox. We imported our data from the programming languages as a CSV file for both 2019 and 2020 into one map in order to pinpoint the popular destinations from the two years. The map produced is color coded in order to differentiate the two years. One can observe that both years have five overlapping popular destinations.

The results we acquired from our research develop several key points to be made. We noticed that five of the top 10 bike stations from 2019 and 2020 remained the same. This makes sense as it establishes a relative consistency in ridership patterns from 2019 and 2020. We originally hypothesized that the majority of the top stations for 2020 would be located around healthcare centers for people to get treatment, but were surprised to find that most stations were located near recreational areas such as parks. A possible explanation for the change in top 10 stations from 2019 to 2020 would be new policies like the stay at home order and social-distancing. A reason why the majority of the popular stations were located near parks would be people’s desire to engage in outdoor activities for entertainment, staying healthy, or escaping the stresses of being stuck at home. It’s worth noting that a station near Memorial Cancer Center was the top 3rd station for 2020, when in 2019 it wasn’t even in top 10. This suggests the growing importance of hospital trips during a pandemic like COVID-19.

In general, our visuals show that ridership around March, 2019 was less compared to ridership in March, 2020. Coming into March 2020, the Citi Bike service had more riders accessing their program, but once COVID-19 struck, the trips plummeted to numbers below 2019’s trip numbers. The greater ridership in the beginning of March 2020 can be attributed to the increasing popularity of the Citi Bike program and in the increasing number of Citi Bike docks, which expands its accessibility through the city.

However, around mid-March of 2020 there was a noticeable drop in ridership. This trend coincides with the closing down of schools in New York City, the issuing of a stay at home order, and the closing of non-essential businesses. Due to this, the number of trips being made significantly dropped as the city was closing down.

Another comparison between 2019 and 2020 was that while 2019’s trips remained relatively high, there were occasional spikes down, but in 2020, the trips remained relatively low, while there were noticeable spikes up. Most of the spikes in the 2019 graph focused around weekdays and cultural holidays like Cinco de Mayo or the start of Ramadan. For 2020, most of the spikes upward occurred around weekends, where people would have the leisure time to go out for bike rides.

From an overall standpoint, ridership has been steadily increasing in 2020, while in 2019 the trend remained relatively constant. The consistency of bike trips in 2019 can be explained by the fact that there weren’t many external stresses limiting bikers. But in 2020, the COVID-19 pandemic had been a severe blow to the Citi Bike program, and so the graph is steadily increasing to mark its recovery from this stress. Additionally, as the months progressed from March to June in 2020, New York has become more lenient with policies, thus allowing people to enjoy increased bike rides outdoors. On top of that, in the beginning of May, the MTA subway system shut down its hours from 1 AM to 5 AM and reopened back to 24 hour service in June. This might be a factor on why ridership in Citi Bike kept increasing in May to June.

Finally, overall ridership for 2020 has significantly decreased. Our time-series visual shows us that for 2019, the number of riders usually ranged anywhere from 0 to 50,000 trips. However, for 2020 that range shrunk from 50,000 trips to 30,000 trips which is approximately a 20,000 difference in trips. Again, this can be explained by the decisions that were made to ensure the safety of New Yorkers. It is also noticed that in our “Monthly Revenue Comparison” visual, revenue in March-May of 2020 was lower than that of March-May 2019. However, as restrictions were beginning to lift, we can see that Citi Bike revenue rebounded in June 2020, making it even higher than revenue in June 2019.

Overall, it can be determined that the results support our hypothesis regarding the negative impact that COVID-19 has had on bike ridership in New York City. The visuals presented display the difference in the number of riders between 2019 and 2020. The data also supports the idea that ridership will rebound from the hit it took, even if it may not reach the number it had in 2019. From the results, it can be noted that 2020 ridership, while not as high, has consistently spiked up as the date progressed to June. For mid-June, it can be concluded that ridership has maintained some sort of consistency. Although we did not research the trends of ridership in 2019, (only displayed for comparison purposes,) there were several steep spikes down. Researching factors that could have contributed to this decline may serve as a valuable point in future research.

We expected a completely different map for 2020 regarding where hotspots were located, however, the top stations did not drastically change from last year. It is noted from the results that 5 stations from the two years overlapped. One noticeable feature about the top 10 stations is that most of these stations were near recreational areas, such as parks.

Finally, we expected most of the popular stations this year to reside near hospitals or other medical centers. However, referring to the point above, only one station was close to a medical center, that being 1 Ave. & E 68 St. We believe that this is because most people would not use bikes in order to get medical attention, but rather personal vehicles.

There are multiple points of interest that have risen while conducting this research. During a time where the way of transportation has been significantly different from our usual way of life, we hope the trends we discovered in this project will help contribute to where and how future bike-sharing systems will be implemented. This project will help station planning for future Citi-Bikes in order to return a higher engagement with the bikes.

When conducting literary research, we have noticed the disparity between station placement throughout the five boroughs. Most stations are located in Manhattan and Brooklyn. Placing new stations in the rest of the boroughs is important because it will provide another form of transportation. Additionally, have bikes at hand encourage social-distancing since each bike is personal to the user.

The methods used in this research project may be replicated for different modes of transportation, such as subway systems. With bike-sharing on the rise, it will be interesting to see how other forms of public transport respond in their effort to bring usage back to normal while ensuring the safety of civilians. There has been research conducted comparing usage of taxis as well. Such research may be extended using similar programs used in this project.

This research project was done with the intent of researching the bike-sharing system in New York City. However, the same can be done for bike-sharing systems throughout the country and around the world. In a broader aspect, this can also be done for biking as a whole. However, this may be different because data may have to be simulated in order to determine biking behavior.

1. DeMaio, P.J. (2003). "Smart bikes: Public transportation for the 21st century." Transportation Quarterly. 57. 9-11. http://www.metrobike.net/wp-content/uploads/2013/10/Smart-Bikes.pdf

2. Dwyer, Matthew. “Testing Trip Duration of Citibike Trips During Commuting Hours.” Authorea, 7 Nov. 2017, www.authorea.com/users/175111/articles/211184-testing-trip-duration-of-citibike-trips-during-commuting-hours.

3. Kanik, Alexandra, and Jacob Serebrin. “The Decisions Cities Made about Coronavirus Had a Big Impact on Bike-Share Ridership.” CityMetric, 21 July 2020, www.citymetric.com/transport/decisions-cities-made-about-coronavirus-had-big-impact-bike-share-ridership-5218.

4. Kaufman, Sarah M., and Jenny O'Connell. “Citi Bike: What Current Use and Activity Suggests for the Future of the Program.” NYU WAGNER, Rudin Center , Feb. 2017, wagner.nyu.edu/files/rudincenter/2017/03/CitiBikeUsage_Mar2017.pdf.

5. Keler, Andreas, et al. “Extracting Commuter-Specific Destination Hotspots from Trip Destination Data – Comparing the Boro Taxi Service with Citi Bike in NYC.” Taylor & Francis, 10 June 2019, www.tandfonline.com/doi/full/10.1080/10095020.2019.1621008.

6. Praveen. “New York City Rides on Citi-Bikes : A Daily Commute or a Fancy Ride .” New York City Rides on Citi-Bikes : A Daily Commute or a Fancy Ride , Authorea, 8 Nov. 2017, www.authorea.com/users/175485/articles/211543-new-york-city-rides-on-citi-bikes-a-daily-commute-or-a-fancy-ride.

7. Pucher, John, and Ralph Buehler. Cycling towards a More Sustainable Transport Future, Taylor & Francis, 19 June 2017, www.tandfonline.com/doi/full/10.1080/01441647.2017.1340234?src=recsys.

8. Teixeira, João Filipe, and Miguel Lopes. “The Link between Bike Sharing and Subway Use during the COVID-19 Pandemic: The Case-Study of New York's Citi Bike.” Transportation Research Interdisciplinary Perspectives, Elsevier, 8 July 2020, www.sciencedirect.com/science/article/pii/S2590198220300774.

9. Citi-Bike Operating Report

10. Citi-Bike System Data

Mentors:

• Suzana Duran Bernardes: sdb425@nyu.edu

• Jingquin Gao: jingquin.gao@nyu.edu

Professor:

• Kaan Ozbay: kaan.ozbay@gmail.com