5 things bike-sharing tells us about Chicago
Divvy bikes are a distinctive trait of the City of Chicago. If you live in the Windy City, you won’t spend a day without spotting at least one of those. With more than 22 million rides, ~650 stations installed, and an average of 8500 rides per day since the program started, these data have stories to tell. We combed through Divvy bikes’ open data and picked the most interesting stories to share via an online platform we developed, so you can see how Chicagoans ride across the city.
Our analysis tells the story of a service that is popular among residents, particularly loved by people in their thirties, and characterized by interesting differences in how people use it.
Curious about what we found out? Here are 5 different stories and don’t forget to play with our data visualizations here.
But before we begin, what is the Divvy Bikes open dataset?
The dataset released by Divvy has a pretty straightforward structure: a list of bike rides with information about where and when a user unlocked and then parked the bike. All the data released by Divvy are anonymized and up until the end of 2019, they also included riders’ age and gender. Our analysis includes data from 2013 until the end of June 2020.
As is typical in any data story, our work started with cleaning the data and preparing them for the analysis. Now, on to the results.
Popular Rides
It is no surprise that the most popular routes span across Downtown and the lovely Chicago Lakefront Trail. In this visualization, you will see the top 1000 rides by popularity. Each arc in the map represents all the trips between the two stations connected.
Looking closer at the map, you will notice two common patterns.
First, several popular rides start or end on university campuses, showing how students use bikes to travel across their campuses.
Second, Divvy bikes seem to be a solution to the “last mile connections” problem. People use Divvy to connect from public transportation to their final destinations, or vice versa when no other public transport is available. You can notice this pattern by looking at the high number of arches connected to the rail lines, like a centipede anchored at the L stops.
Popular Stations
As expected, the most used L stations are strongly connected to the most popular rides. However, while popular rides span across a large area of Chicago, popular stations are densely concentrated around Downtown and the Northside, peaking near Union Station and the Navy Pier. By the way, did you know that Union Station connects 140,000 passengers with the City on an average weekday?
Gender Disparity
This visualization has probably been the most surprising for us. As you can see, the whole Divvy’s service area is blueish and that is because 75% of rides are done by males. Among 600+ stations, only 14 have women as the majority of their users. The trend seems to span the city without exceptions.
Looking at the map below, you won’t see a single neighborhood colored in pink.
Rider’s age
The average age of a Divvy rider is between 30 and 40 years old. However, this seems to be true only for the most popular areas.
Outside Downtown and the North Side, users tend to have a different age distribution, with the average age shifted older. Nevertheless, some stations across the maps show a highly biased usage towards the younger population. Higher and reddish columns in the map visualization represent all these atypical stations.
Newest Stations
In this view you will see the station’s age: bigger and greener dots represent newer stations.
Since the program started in 2013, Divvy has constantly improved its coverage, adding stations year by year. In particular, after 2015, Divvy has been expanding into the West and the South sides of Chicagoland. Starting in the summer of 2020, they started offering e-bikes in these areas without additional fees. Well done Divvy!
Conclusions
As data scientists, part of what we do is looking for hidden patterns and insights in the data. Why are women using bike-sharing less than men? Why does the average age change so much outside Downtown? These are just some of the questions that we asked ourselves looking at the results of our analysis. We’d love to collaborate with city planners and communities to find answers and, possibly, improve our city’s bike-sharing service and public transportation. In the meantime, we hope that this article will continue to ignite many other interesting questions and conversations.
Being involved in projects where we can use data science to make our communities better is something that we love doing at Wepo. For example, during the United States 2020 Census, we provided the Illinois Department of Human Services and the University of Illinois at Chicago with data science services to coordinate the outreach efforts of more than 360 local organizations.
If you want to know more about Wepo, visit wepo.io where you can find our contacts. We would love to connect with you!
See you on your next ride!
Tools
Data Science: Python, Pandas, Geopandas, PyDeck
Web Development: React, DeckGL, Flask, FastAPI
Databases: MongoDB, Postgres
Deployment: Docker, Kubernetes
Cloud Provider: Google Cloud Platform, Mongo Atlas
