The purpose of this project was to investigate the impact of the COVID-19 pandemic on New York City's Citibike ride sharing service. In order to do this, I aggregated data from the Citi Bike Data webpage from January 2020 through November 2020, cleaned and analyzed the data for roughly 19 million rides, and then built a Tableau Story to provide a narrative that would help explain the impact of the pandemic on the Citibike program. The Tableau Story includes two interactive dashboards so that users can further explore the data.
This Belly Button Plot.ly project creates an interactive dashboard that allows the user to analyze the microbial portrait of each test's subject navel (belly button). In 2011, the Public Science Lab at NC State began a study into the microbes that inhabit the human belly button, and the data supplied for this dashboard came directly out of that study.
In order to view the Tableau story and dashboards, you can either download the 'citibike.jl.twb' file in this repo and view in your Tableau desktop app or you can also access it online by visiting my Tableau Public account. In the top right-hand corner of all the views, there are sliders that allow you to view the entire eleven month timeframe or drill down into specific months.
In the GitHub repository for this project you will find all the resources necessary to view this Tableau story and dashboard. Here's a quick rundow of those files and a brief explanation:
- citibike.jl.twb - the Tableau workbook file that contains the full Citibike story, plus the dashboards, and the worksheets that were built for the creation of the Citibike story.
- concat_csv.ipynb - the Jupyter notebook file that used to combine all of the NY & Jersey City (JC) files that were downloaded as CSV files from the Citibike website.
- in the resources folder you will find all the csv files for ONLY the JC trips. The NYC trip files are to large right now to upload into this repository.
- the image folder only contains screenshots that were used to create the README file
It comes as no surprise to report that the Covid19 pandemic had significant impact on CitiBike ridership during 2020. Prior to downloading and analyzing the data, I knew the effects of the pandemic would have a negative impact on the program--but I was unsure as to what the final analysis would show. Here are the main three phenomena that are most important and deserve to be pointed out:
#1 -- When the lockdown began in late March, as expected, the number of total rides decreased dramatically--even as the weather was warming and becoming more conducive to bike travel. What wasnt't expected was the dramatic shift away from the typically busiest start stations in Mid-Town, Chelsea, NYU, etc to a much more decentralized dispersement.
#2--The second phenomenon that was discovered via analysis of the data, which was maybe a little bit easier to anticipate, was the big decline we saw in the 8am hour for starttimes. With lockdown causing businesses and schools to close, we see how important those morning rush hour riders are to our overall numbers. As restrictions started to ease and re-opening phases began--we see overall ridership increase, especally for that morning rush hour. We should consider incentivizing those morning rush hour riders to come back now that they've maybe they're still not having to go into the office or classroom, or have found alternate methods for getting to where their going in the morning. At the very least we should do a campaign educating the public about how the bikes are cleaned and that they are safe.
#3--The Final phenomenon to address is that we are still having a significant percentage of riders fall into the "unknown" gender category. Since we're always striving for better user data so that we can more precisely tailor our maketing campaings, I think we should institute so minor changes to see if we can convert some of those "unknown" riders into a category. And what I am suggesting is to increase the number of "gender" categories from just two (male & female). Some riders may feel those categories too gender binary and in an effort to be more inclusive we could add several gender-neutral designations. This is a topic that could be further investigated in order to provide some good options to be added to the possible selections.
John Loucks
Email: [email protected]
GitHub: https://github.com/loucksjohn
link to tableau public: https://public.tableau.com/profile/john.loucks5224#!/vizhome/CitiBikeStory_16116200207080/CitiBikeStory