Cyclistic bike‑share analysis

Google Data Analytics capstone, June 2022

R, tidyverse, ggplot2

At a glance

  • What: a course exercise for a bike-share company that wants more riders on single-ride or day passes (the course calls them casual riders) to buy yearly memberships, asking how the two groups ride differently.
  • My part: I did the analysis in R and wrote the report with recommendations; it was an individual course capstone.
  • Result: casual riders take longer rides, a median of 16 minutes against 9, and ride more at weekends; members ride more on weekdays, with a morning peak at 8 am.

The question

In the Google course case study, Cyclistic wants to turn more casual riders, who buy single-ride or day passes, into yearly members. The question was how the two groups use the bikes differently.

Data and method

I used twelve monthly trip files from Divvy, Chicago's bike-share system, from June 2021 to May 2022: 5,860,776 rides. In R I removed three kinds of ride: rides with a negative ride time, rides from two test or warehouse stations, and rides with any missing value. Almost all of the last kind were e-bike rides with no station recorded. That left 4,666,342 rides. I then compared the two groups by ride length, day, hour, month and route, and drew the charts with ggplot2.

What I found

Of the rides kept, members took 57% and casual riders 43%. Casual rides ran longer: 8.9% lasted over an hour against 0.6% for members. They were also more often at weekends: 41% of casual rides against 25% of member rides. Round trips make up 6 of their top 10 routes, although only about 1 in 10 casual rides is a round trip. Members' morning peak at 8 am suggests many of them ride to work, and both groups ride far more in summer.

Line chart of rides per month for casual riders and members, June 2021 to May 2022: highest in summer, lowest in January.
Rides per month, June 2021 to May 2022. Casual rides peak in July, member rides in August.
Line chart of rides by hour of the day: both groups peak at 5 pm, members also peak at 8 am.
Rides by hour of the day. Both groups peak at 5 pm; members also peak at 8 am.
Bar chart of the 10 most used routes by casual riders, led by the round trip from Streeter Dr and Grand Ave.
Top 10 routes for casual riders. Six start and end at the same station.

Recommendations

In my report I recommended putting more bikes and marketing on the routes casual riders use most, offering a weekend pass, selling passes by number of rides, such as 5 or 10, to regular casual riders, and running campaigns in the warmer months.

What I would do next

I would recommend better data collection, with fewer missing values, negative ride times and repair trips in the data. I would also use more than one year of data, and add surveys or interviews to learn what casual riders want.