R Notebook

Problem

Business Tasks

Cyclistic is a fictional bike-share company inside Chicago. This company tasked us with the analysis of this question:
How do annual members and casual riders use Cyclistic bikes differently?
The “casual” riders are the clients that use single-ride and full day pass. Those offer are about short term engagement.
The “member” riders are the clients that buy annual membership. This offer is about long term engagement.
In order to convert “casual” clients into “member” clients, we could try to develop new “transition” offer that is about mid term engagement. This offer would be developed to be attractive to “casual” clients while being unattractive to “member” clients since we don’t want to convert members into casuals. In order to find opportunities for said new transition offers, we will analyse the riders behaviors.
Let’s break this business task into sub tasks:

Identify and categorize how users use Cyclistic services
See how those usage relate to casual riders and annual members
Identify usage differences and similarities between casual riders and annual members

Audience (stakeholders in bold):

Director of marketing
Part of analytics team
Cyclistic executives

Data sources:

Cyclistic’s first-party data from the last 12 months: from march 22 to February 23. Cyclistic is a fictional company so the data is provided by Motivate International Inc. The data contain one row per user’s trip. The variables are about:
- trip’s starting hour and location
- trip’s ending hour and location
- customer’s membership
- bike type
The files used are:
- 2022MM-divvy-tripdata/2022MM-divvy-tripdata.csv with MM the month (03 to 12)
- 202209-divvy-tripdata/202209-divvy-publictripdata.csv
- 2022MM-divvy-tripdata/2022MM-divvy-tripdata.csv with MM the month (01 to 02)
Chicago ZIP boundaries Contains a shapefile map that represents the Chicago divided by ZIP areas

dataPath <- 'absolutePath'
knitr::opts_knit$set(root.dir = normalizePath(dataPath))

Analysis

Data cleaning

Removed data:

duplicate trips
the trips that last for a minute or less because it’s a canceled trip
the trips that last for more than a day because it’s likely that the user forgot to return the bike
the trips that had starting or ending location outside of Chicago because that’s not supposed to be possible
the trips with a duration that are in the 1% longest trips to make readable the graphs that have tripduration as an axis

Data modifications:

added trip duration
deteriorated the coordinates so that the possible number of coordinates inside Chicago is finite
added day of the week
added month
added ZIP zone IDs for the trips starting location and ending location
ordered by station name and then by start time

Usage by time

Usage duration

Both casual and member clients trip’s duration is around 5 min but members are more concentrated around that duration.

Usage throughout the day

Casuals use the service all along the day, with a cliff around 5:30 AM and a peak around 5:30 PM
Member’s graph is similar casual’s graph with two variations:

another peak around 8:00 AM
less usage during working hours

Usage throughout the week

Casuals use Cyclistic a lot more during the weekend than during the workweek. On the opposite, members have a close to constant usage during the whole week but they use the service slightly more during the workweek than during the weekend.

Usage throughout the year

Casuals enjoy Cyclistic during the sunny months (more than 15% usage during June to august) and avoid it during the cold ones (less than 2.5% usage during the December to February).
Members follow the same trend, but their usage is much more consistant, with usage under 4% only in December and no month with usage above 12.5%.

Recommendations

The analysis aims at uncovering the differences between “casual” and “members” clients in order to understand what could convince a “casual” client to become a “member” client.
To achieve this goal, we will make 3 recommendations that aims at producing transitional offer for casuals toward annual pass or making the annual pass more interesting for casual. Those recommendations are:

Creation of a summer pass since it’s the season that interest casual the most. The member shall keep it’s current offer to have access to the service during the whole year.
Creation of an annual leisure pass that gives:
- full access during weekends
- access between 19h-5h during workweek
  
  Those time ranges interest the casuals. The member shall keep his current offer to have access to the two commuting time range.
On annual passes, allow the member to loan a 2nd bike for free ~3 times per week. This could make long time offers more attractive to casuals and at the same times it is a way for non customers to test the service in very favorable conditions (social usage).

Appendix

Unused findings

Usage by location

Location alone

At first glance, both members and casuals make a similar usage of Cyclistic in term of location alone: they are both centered on the Lincoln park area of ZIP code 60614. However, by analyzing the difference between member usage we can see that:

casuals prefer Streeterville neighborhood, Lincoln park and South Loop a bit more
members prefer West Loop and Woodlawn a bit more

These tendencies remain the same during the week, but get reinforced for Streeterville during the weekend.
We only display the graphs based on the trip ending zone because the graph based on trip beginning zone are similar.

## [[1]]

## 
## [[2]]

Location during workweek

Casuals are slightly more concentrated on Lincoln Park than members during the workweek

## [[1]]

## 
## [[2]]

Location during weekend

Members are slightly more concentrated on Lincoln Park during the weekend

## [[1]]

## 
## [[2]]