A year-long analysis of Chicago's Cyclistic bike-share program comparing casual riders and annual members. This capstone project follows the full Google Data Analytics process to inform marketing strategies that convert high-usage casual riders into long-term members.
Project overview
Combined 12 months of Cyclistic trip data from Chicago to uncover behavioral differences between casual riders and annual members. Using R (tidyverse, lubridate, janitor, ggplot2) for preparation and analysis, plus Tableau for visualization, explored ride volume, duration and timing patterns across seasons and weekdays.
Key deliverables
Phase 1
Import & Standardization
Downloaded and combined 12 monthly CSVs using read_csv(), map_df() and clean_names().
Phase 2
Cleaning
Removed invalid rides, parsed timestamps and dropped rows with missing data.
Phase 3
Feature engineering
Engineered ride length, weekday labels and features for behavioral segmentation.
Phase 4
Analysis
Summarized rides by rider type, weekday and season to reveal core patterns.
Phase 5
Visualization
Created ggplot2 charts and Tableau dashboard for stakeholder storytelling.
The project begins by downloading 12 monthly CSVs from the Divvy/Cyclistic portal. Using tidyverse, these files were imported and combined into a single, clean dataframe ready for analysis.
RStudio was configured with tidyverse, lubridate, janitor, ggplot2 and readr to support the full workflow from import through visualization.
RStudio console confirming successful installation of required packages.
Using list.files(), map_df() and clean_names(), the 12 monthly CSVs were imported and standardized into one dataframe with normalized column names.
RStudio environment after successfully importing and combining 12 months of trip data.
After import, focus shifted to removing invalid records and creating features that explain rider behavior. Using lubridate and dplyr, ride length and weekday labels were built.
Start and end timestamps were parsed into datetime objects, then used to calculate ride duration in minutes. Invalid rides were removed to avoid skewed averages.
RStudio console showing mutation steps for ride length and weekday plus filters for invalid rides.
To support segmentation, day_of_week labels were created and used to compare weekday commuting patterns with weekend leisure trips across rider types.
Feature engineering snippet adding ride length and weekday labels to the combined dataset.
With a clean dataset, behavior was summarized by rider type, weekday and season. These statistics reveal that casual riders take longer trips concentrated on weekends while members ride more on weekdays.
Using group_by() and summarise(), mean, median, min and max ride length were computed for each rider segment along with total ride counts.
Summary statistics showing ride length and volume for casual and member riders.
Additional aggregations by weekday and month surfaced clear seasonality: casual riders cluster in warmer months and on weekends, while member activity resembles workweek commute patterns.
Rides by day of week: casual volume spikes on weekends, member volume peaks Monday-Friday.
Visualization combines R-based exploration with a polished Tableau dashboard. ggplot2 charts help iterate on questions while Tableau delivers an interactive view for stakeholders.
In R, bar charts were created comparing ride counts and average duration by weekday and rider type. These plots validate behavioral patterns before investing in dashboard design.
Average ride duration by day of week showing longer casual rides on weekends.
Average ride duration by rider type: casual riders stay out nearly twice as long as members.
These snippets show how the analysis flows from package setup through import, cleaning, summary statistics and plotting. They mirror the structure of the Cyclistic case study.
Package setup and import
install.packages(c("tidyverse","lubridate","janitor"))
library(tidyverse)
library(lubridate)
library(janitor)
file_list <- list.files("data_raw", pattern = "*.csv")
all_trips <- file_list %>%
map_df(read_csv) %>%
clean_names()
Cleaning and feature engineering
all_trips <- all_trips %>%
mutate(
ride_length = difftime(ended_at, started_at),
day_of_week = wday(started_at, label = TRUE)
) %>%
filter(ride_length > 1) %>%
drop_na()
summary_stats <- all_trips %>%
group_by(member_casual) %>%
summarise(avg = mean(ride_length))
The final phase packages the cleaned dataset for Tableau, where stakeholders can explore rider behavior, filter by segment and drill into KPIs that tie directly to membership growth strategy.
The Tableau dashboard summarizes 12 months of rider behavior, comparing casual and member riders with filters for rider type, day of week and hour of day.
Tableau dashboard combining ride volume, duration and seasonal trends to support membership decisions.
The final dashboard delivers a consolidated view of Cyclistic usage, highlighting how casual riders and members differ by volume and duration. The metrics below summarize the year-long dataset.
4.17M
Total rides: combined casual and member trips after cleaning.
2.32M
Member rides: volume associated with annual subscribers.
1.85M
Casual rides: pay-per-ride and day pass usage.
26.7
Avg duration casual: minutes per trip for casual riders.
12.5
Avg duration member: minutes per trip for members.
Visual summary of Cyclistic rider behavior split by rider type, weekday and season.
Weekday distribution showing casual weekend peaks and member weekday consistency.
Casual usage spikes on weekends and in warmer months, suggesting leisure or tourism service use.
Members ride consistently on weekdays with shorter, repeatable trips aligning with commuting.
Casual riders spend nearly twice as long per ride as members, indicating different use cases.
High-frequency casual riders could be targeted with tailored membership offers.
Promotions timed to peak casual months can nudge repeat users toward memberships.
The dashboard turns millions of records into a simple story supporting marketing decisions.
Objective: Convert high-usage casual riders into entry-level members.
Objective: Keep members engaged through consistent value and convenience.
Objective: Use behavioral data to keep the system efficient and rider-friendly.
This Cyclistic rider analysis connects raw open trip data to concrete membership strategy. By importing and cleaning 12 months of rides in R, engineering features for segmentation, and surfacing patterns in Tableau, the project delivers a clear picture of when and how different riders use the system — creating a natural path for targeted conversions, retention campaigns and operations planning grounded in real behavior.