Cyclistic: a bike-share program

This is a case study for the Google Data Analytics Certificate. For complete instructions about this case study, check here

In this case study, I will follow the steps of the Data Analysis Process:

1. Ask
2. Prepare
3. Process
4. Analyze
5. Share
6. Act

Step 1: Ask

The business task:

I analyze Cyclistic's historical bike trip data from April 2020 to March 2021 to identify trends on how members and casual riders use Cyclistic bikes differently; Then, I generate 3 recommendations based on this analysis for the Cyclistic's marketing team to design marketing strategies for converting casual riders into members.

Step 2: Prepare

Note: For more details of this part, check this Notebook here.

The data I used in this case study can be downloaded from here.

• Note: the datasets have a different name because Cyclistic is a fictional company. For the purpose of this case study, the datasets are appropriate and will enable to answer the business questions.

The data contains 12 datasets in total:

Each dataset contains bike trip data of a month from April 2020 to March 2021, and it is consisted of 13 columns.

After examining the data, I believe the data meets the ROCCC standard:

The data is reliable, original, comprehensive, current, and cited:

• it has been made available under this license by Motivate International Inc with permissions from the City of Chicago's Divvy bicycle sharing service;
• it is owned by the City of Chicago;
• it is updated by Motivate International Inc on a monthly basis.

The potential problems with the data:

The data is provided "as-is" by Motivated International Inc. Therefore, there are certain amount of invalid data in the data. As a result, the data requires some cleaning and transformation before it can be applied for further analysis.

Step 3: Process

Note: For more details of this part, check this Notebook here.

These original datasets are downloaded and stored:

• on Google Drive and locally on my computer

For each dataset, I perform these data cleaning and transformations:

• check the uniqueness of ride_id
• drop any null values in all columns
• check the data types of each column, and convert to the correct type if necessary
• check the consistency between start_station_name and start_station_id
• check the consistency between end_station_name and end_station_id
• drop irrelevant columns: start_lat, start_lng, end_lat, end_lng

Step 4: Analyze

Note: For more details of this part, check this Notebook here

After cleaning the original datasets, I store them on:

• Google Drive
• locally on my computer

Then, I concatenated these datasets into one dataset named btrips, and divide this dataframe into two sub-dataframes:

• m_btrips for members
• c_btrips for casual riders

Next, I perform these analysis using these two sub-dataframes:

• calculate the amount of members vs. casual riders have used Cyclistic bikes
• calculate the amount of members vs. casual riders used Cyclistic bikes by months
• discover the type of bikes members vs. casual riders perfer
• calculate the length of bike rides members vs. casual riders use
• discover which day in a week that members vs. casual riders start their bike rides
• calculate the average length of bike rides members vs. casual riders use based on days in a week
• discover the top 5 bike stations members vs. casual riders visited most frequently to start their bike rides

Finally, I identify these trends of how members and casual riders use Cyclistic bikes differently:

• 41% Cyclistic users are casual riders.
• The amount of Cyclistic bikes used by members and casual riders has been declining between August to January.
• 82% casual riders have used docked bikes.
• For bike rides within 12 hours, the amount of casual riders are more than members; For bike rides more than 12 hours, the amount of members are significantly more than casual riders.
• On weekdays, the amount of members use Cyclistic bikes are significantly more than casual riders, however, the amount gap is less wide on weekends.
• The average length of bike rides of casual riders is less than 20 minutes; Meanwhile, the average length of bike rides of members is over 40 minutes.
• The top 5 bike stations casual riders start their bike rides are different from the top 5 stations members usually start their bike rides.

Step 5: Share

Note: For more details of this part, check this PowerPoint here

Based on my analysis, I summarize that annaul members and casual riders use Cyclistic bikes differently via:

• seasons
• bike types
• bike ride lengths
• weekdays vs. weekends
• bike stations

Step 6: Act

Note: For more details of this part, check this PowerPoint here

The final conclusion based on my analysis:

From April 2020 to March 2021, 41% of Cyclistic bike users were casual riders. Therefore, it is essential for Cyclistic to design marketing strategies targeted at maximally converting casual riders into annual members.

Therefore, I generate 3 recommendations for the Marketing Team:

1. Offer fare discounts to casual riders using Cyclistic bikes shorter than 12 hours for each ride and/or on weekends.
2. Design promotions specifically targeted at casual riders between February to July.
3. Launch digital and non-digital ads at the 5 bike stations that casual riders used most, and put more docked bikes at these stations.

Next steps to take:

• Develop a deeper research on why and why not casual riders would buy Cyclistic annual memberships
• Generate a comprehensive analysis on how to use digital media to influence casual riders to become members

Additional data I could use to expand my findings:

• Cyclistic's historical bike fare data
• data from other peer competitors in the Chicago area; data should includes information such as their bike fares charged to members versus non-members and the amount of members purchased their annual memberships.