Divvy Bike Share


This is a case study of twelve months of bike ride data from Divvy, a bike share service in Chicago, Illinois.

The bike ride data was analyzed using Excel, R, and Tableau to determine the key differences between annual Divvy members and casual bike riders.


The previous twelve months of bike ride data was downloaded from Divvy, see https://divvy-tripdata.s3.amazonaws.com/index.html. This dataset includes the information of over 5.8 million bike rides through the Divvy bike share service, ranging from May 2022 to April 2023.  The information gathered for each ride includes the type of bike used, the start and end times, the start and end bike stations, and the membership status of the user.

I first used Excel to clean each of the datasets.  By using the start and end times of each ride, I calculated the duration of each ride in the dataset.  I also calculated the day of the week of each ride using Excel's WEEKDAY() function.

For the actual analysis, I used R due to the large size of the dataset and the limitations of Excel.  See my R Markdown notebook for details of the steps taken in R to analyze the data.  First, I combined the 12 datasets into one and removed some of the columns I wouldn't need for my analysis.  I also extracted the months, days and years of each bike ride from the data.  I used the Tidyverse and some of its packages, such as Tidyr and Dplyr, in order to clean and analyze the data.  I ordered the datasets by day of the week and by month.  I also aggregated the data by membership type in order to see the trends between annual members and casual users.  The main insights I was looking for were the average ride times and number of rides taken by members and casual users.  I looked for trends on a weekday basis, and by a month-to-month basis.  After concluding the analysis in R, I exported my findings to Tableau in order to visualize the data.

In Tableau, I created a dashboard to visualize the bike ride data.  The dataset was split between annual members and casual riders to highlight their differences.  There were three main points to compare the members and casual riders: the type of bike ridden, the day of the week, and the month.  For each of these points, I plotted the total number of bike rides for each group and also the average ride duration for each group.


The data provided a number of insights into the usage patterns of Divvy's annual members and their casual users.  Annual members did not use Divvy's docked bikes, instead using classic and electric bikes evenly.  They also used these bikes consistently throughout the week.  Furthermore, they have a drop-off in rides taken during winter months.  Casual riders, on the other hand, use all 3 types of bikes.  They use electric bikes the most and docked bikes the least.  Unlike annual members, casual riders are not consistent in their bike rides, leading to weekends having the lowest numbers in the dataset.  However, like annual members, casual riders also have a drastic drop-off in rides in the winter months.  But in these months, casual rides fall much harder, falling as low as 40,000 rides in January 2023, compared to the 150,000 rides from annual members in the same month.  Surprisingly, the trip duration data was eye-opening.  On average, casual riders rode their Divvy bikes over four times longer than annual members.  This leads me to believe that annual members use the bikes differently, perhaps for their daily commute through the city, while casual rides were for leisure.


An R Markdown notebook detailing my analysis process in R.

An interactive Tableau dashboard to browse the visualized data.

A presentation to display my findings.