Cyclistic Bike-Share Case Study
Google Data Analytics Capstone Project
Bethany Leach
2023-06-13
Scenario
You are a junior data analyst working in the marketing analyst team at Cyclistic, a bike-share company in Chicago. The director of marketing believes the company’s future success depends on maximizing the number of annual memberships. (Therefore, your team wants to understand how casual riders and annual members use Cyclistic bikes differently). From these insights, your team will design a new marketing strategy to convert casual riders into annual members. But first, Cyclistic executives must approve your recommendations, so they must be backed up with compelling data insights and professional data visualizations.
This course included six steps to the data analysis process, which will be outlined in this project.
- Data Analysis Process
- Ask
- Prepare
- Process
- Analyze
- Share
- Act
Data Analysis Process
Step 1 - Ask
- Guiding Questions
- How do annual members and casual riders use Cyclistic bikes differently?
- Why would casual riders buy Cyclistic annual memberships?
- How can Cyclistic use digital media to influence casual riders to become members?
- Understanding Key Stakeholders and Their
Expectations
- Cyclistic Executive Team - The notoriously detail-oriented executive team will decide whether to approve the recommended marketing program.
- Lily Moreno - The director of marketing and your manager. Moreno is responsible for the development of campaigns and initiatives to promote the bike-share program. These may include email, social media, and other channels.
- Define the Problem You’re Trying to Solve
- Cyclistic’s financial analysts have concluded that annual members are much more profitable than casual riders. Moreno believe’s that Cyclistic’s future growth is intrinsically tied to maximizing the number of annual members. Therefore, the business task here is to design marketing strategies aimed at converting casual riderrs into annual members.
Step 2 - Prepare
- Downloaded data per instructions in Case Study 1
- Downloaded 12 months worth of ride data from (https://divvy-tripdata.s3.amazonaws.com/index.html).
- 12 .csv files were downloaded (4/22-3/23) and were saved locally before importing the files into excel and then later, RStudio
- Database Structures
- Cyclistic: Previous 12 months (4/22-3/23) of Bike Share data is organized into 12 different .csv files. Each csv file has 13 different columns with a varying number of data types.
- Addressing Issues of Bias and Credibility
- The data has been made available by Motivate International Inc. under this license
- This is a public dataset that abides by data-privacy by not providing riders’ personally identifiable information such as credit card numbers.
Step 3 - Process
I used excel to preview the data initially. I did some preliminary cleaning of the data in excel and added the ride_length column. I used R to organize, clean, and process the data. I then used R to create data visualizations. I couldn’t use Tableau public since my dataset (~6.1million entries) was too large.
- Excel
- I added a column to each of the 12 .csv files titled ride_length in the format HH:MM:SS = (2400 - start_time) + end_time
- I added a column to each of the 12 .csv files titled day_of_week and calculated the day of the week the each ride began using excel’s WEEKDAY function =WEEKDAY(cell, 1) where 1 refers to Sunday being the first day of the week.
- R - RStudio
- Installed all pertinent packages for exploration in R.
install.packages("readxl", repos = "http://cran.us.r-project.org")
install.packages("dplyr", repos = "http://cran.us.r-project.org")
install.packages("readr", repos = "http://cran.us.r-project.org")
install.packages("janitor", repos = "http://cran.us.r-project.org")
install.packages("lubridate", repos = "http://cran.us.r-project.org")
install.packages("tidyverse", repos = "http://cran.us.r-project.org")
install.packages("leaflet", repos = "http://cran.us.r-project.org")
library(readxl)
library(dplyr)
library(readr)
library(janitor)
library(lubridate)
library(tidyverse)
library(leaflet)- Imported all necessary datasets.
april_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/april_2022_tripdata_revised.csv")
may_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/may_2022_tripdata_revised.csv")
june_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/june_2022_tripdata_revised.csv")
july_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/july_2022_tripdata_revised.csv")
august_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/august_2022_tripdata_revised.csv")
september_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/september_2022_tripdata_revised.csv")
october_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/october_2022_tripdata_revised.csv")
november_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/november_2022_tripdata_revised.csv")
december_22_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/december_2022_tripdata_revised.csv")
january_23_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/january_2023_tripdata_revised.csv")
february_23_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/february_2023_tripdata_revised.csv")
march_23_tripdata <- read_csv("/Users/bethanyleach/Desktop/12_months_cycle_data/march_2023_tripdata_revised.csv")- I ran into parsing issue warning messages while importing the 12 .csv files. This was largely in part due to columns 22-30 on each page of the worksheet. In that range of columns, I found the average and maximum ride lengths as well as the day of the week where the most rides occurred (mode). In this region, I also constructed tables looking at average ride_length for casual users vs members on a daily basis.
- I wrangled the 12 .csv files and combined into a single data frame cyclistic_one_yr_data using the bind_rows function from the dplyr package.
- I then counted the total number of rideable_type and member_casual.
cyclistic_one_yr_data <- bind_rows(april_22_tripdata, may_22_tripdata,
june_22_tripdata, july_22_tripdata,
august_22_tripdata, september_22_tripdata, october_22_tripdata,
november_22_tripdata, december_22_tripdata, january_23_tripdata,
february_23_tripdata, march_23_tripdata)
glimpse(cyclistic_one_yr_data)## Rows: 5,803,720
## Columns: 31
## $ ride_id <chr> "3564070EEFD12711", "0B820C7FCF22F489", "89EEEE3229…
## $ rideable_type <chr> "electric_bike", "classic_bike", "classic_bike", "c…
## $ started_at <chr> "4/6/22 17:42:48", "4/24/22 19:23:07", "4/20/22 19:…
## $ start_date <chr> "4/6/22", "4/24/22", "4/20/22", "4/22/22", "4/16/22…
## $ start_time <time> 17:42:48, 19:23:07, 19:29:08, 21:14:06, 15:56:30, …
## $ ended_at <chr> "4/6/22 17:54:36", "4/24/22 19:43:17", "4/20/22 19:…
## $ end_date <chr> "4/6/22", "4/24/22", "4/20/22", "4/22/22", "4/16/22…
## $ end_time <time> 17:54:36, 19:43:17, 19:35:16, 21:23:29, 16:02:11, …
## $ start_station_name <chr> "Paulina St & Howard St", "Wentworth Ave & Cermak R…
## $ start_station_id <chr> "515", "13075", "TA1307000121", "13075", "TA1307000…
## $ end_station_name <chr> "University Library (NU)", "Green St & Madison St",…
## $ end_station_id <chr> "605", "TA1307000120", "TA1307000120", "KA170600500…
## $ start_lat <dbl> 42.01913, 41.85308, 41.87184, 41.85308, 41.87181, 4…
## $ start_lng <dbl> -87.67353, -87.63193, -87.64664, -87.63193, -87.646…
## $ end_lat <dbl> 42.05294, 41.88189, 41.88189, 41.86749, 41.88224, 4…
## $ end_lng <dbl> -87.67345, -87.64879, -87.64879, -87.63219, -87.641…
## $ member_casual <chr> "member", "member", "member", "casual", "member", "…
## $ ride_length <time> 00:11:48, 00:20:10, 00:06:08, 00:09:23, 00:05:41, …
## $ day_of_week <dbl> 4, 1, 4, 6, 7, 5, 2, 3, 6, 6, 7, 4, 4, 7, 4, 2, 2, …
## $ ...20 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...21 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...22 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...23 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...24 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...25 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...26 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...27 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...28 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...29 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...30 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ ...31 <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…summary(cyclistic_one_yr_data)## ride_id rideable_type started_at start_date
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## start_time ended_at end_date end_time
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class1:hms Class :character Class :character Class1:hms
## Class2:difftime Mode :character Mode :character Class2:difftime
## Mode :numeric Mode :numeric
##
##
##
## start_station_name start_station_id end_station_name end_station_id
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## start_lat start_lng end_lat end_lng
## Min. :41.64 Min. :-87.84 Min. : 0.00 Min. :-88.14
## 1st Qu.:41.88 1st Qu.:-87.66 1st Qu.:41.88 1st Qu.:-87.66
## Median :41.90 Median :-87.64 Median :41.90 Median :-87.64
## Mean :41.90 Mean :-87.65 Mean :41.90 Mean :-87.65
## 3rd Qu.:41.93 3rd Qu.:-87.63 3rd Qu.:41.93 3rd Qu.:-87.63
## Max. :42.07 Max. :-87.52 Max. :42.37 Max. : 0.00
## NA's :5855 NA's :5855
## member_casual ride_length day_of_week ...20
## Length:5803720 Length:5803720 Min. :1.000 Mode:logical
## Class :character Class1:hms 1st Qu.:2.000 NA's:5803720
## Mode :character Class2:difftime Median :4.000
## Mode :numeric Mean :4.103
## 3rd Qu.:6.000
## Max. :7.000
##
## ...21 ...22 ...23 ...24 ...25
## Mode:logical Mode:logical Mode:logical Mode:logical Mode:logical
## NA's:5803720 TRUE:1 TRUE:11 NA's:5803720 NA's:5803720
## NA's:5803719 NA's:5803709
##
##
##
##
## ...26 ...27 ...28 ...29 ...30
## Mode:logical Mode:logical Mode:logical Mode:logical Mode:logical
## NA's:5803720 NA's:5803720 NA's:5803720 NA's:5803720 NA's:5803720
##
##
##
##
##
## ...31
## Mode:logical
## NA's:5803720
##
##
##
##
## cyclistic_one_yr_data_2 <- cyclistic_one_yr_data
count(cyclistic_one_yr_data_2, rideable_type)## # A tibble: 3 × 2
## rideable_type n
## <chr> <int>
## 1 classic_bike 2639559
## 2 docked_bike 173747
## 3 electric_bike 2990414count(cyclistic_one_yr_data_2, member_casual)## # A tibble: 2 × 2
## member_casual n
## <chr> <int>
## 1 casual 2337439
## 2 member 3466281- Since I imported these 12 .csv files after I had performed preliminary calculations in Excel, I dropped the columns of the dataset that would not be necessary for further analysis in R and saved this as a new data frame cyclistic_one_yr_data_3.
cyclistic_one_yr_data_3 <- cyclistic_one_yr_data_2
cyclistic_one_yr_data_3 <- cyclistic_one_yr_data_3 %>%
select(-c(...20,...21,...22,...23,...24,...25,...26,...27,...28,...29,...30,
...31))Step 4 - Analyze
colnames(cyclistic_one_yr_data_3)## [1] "ride_id" "rideable_type" "started_at"
## [4] "start_date" "start_time" "ended_at"
## [7] "end_date" "end_time" "start_station_name"
## [10] "start_station_id" "end_station_name" "end_station_id"
## [13] "start_lat" "start_lng" "end_lat"
## [16] "end_lng" "member_casual" "ride_length"
## [19] "day_of_week"nrow(cyclistic_one_yr_data_3)## [1] 5803720dim(cyclistic_one_yr_data_3)## [1] 5803720 19head(cyclistic_one_yr_data_3)## # A tibble: 6 × 19
## ride_id rideable_type started_at start_date start_time ended_at end_date
## <chr> <chr> <chr> <chr> <time> <chr> <chr>
## 1 3564070EEFD1… electric_bike 4/6/22 17… 4/6/22 17:42:48 4/6/22 … 4/6/22
## 2 0B820C7FCF22… classic_bike 4/24/22 1… 4/24/22 19:23:07 4/24/22… 4/24/22
## 3 89EEEE32293F… classic_bike 4/20/22 1… 4/20/22 19:29:08 4/20/22… 4/20/22
## 4 84D4751AEB31… classic_bike 4/22/22 2… 4/22/22 21:14:06 4/22/22… 4/22/22
## 5 5664BCF0D1DE… electric_bike 4/16/22 1… 4/16/22 15:56:30 4/16/22… 4/16/22
## 6 AA9EB7BD2E1F… classic_bike 4/21/22 1… 4/21/22 16:52:33 4/21/22… 4/21/22
## # ℹ 12 more variables: end_time <time>, start_station_name <chr>,
## # start_station_id <chr>, end_station_name <chr>, end_station_id <chr>,
## # start_lat <dbl>, start_lng <dbl>, end_lat <dbl>, end_lng <dbl>,
## # member_casual <chr>, ride_length <time>, day_of_week <dbl>summary(cyclistic_one_yr_data_3)## ride_id rideable_type started_at start_date
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## start_time ended_at end_date end_time
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class1:hms Class :character Class :character Class1:hms
## Class2:difftime Mode :character Mode :character Class2:difftime
## Mode :numeric Mode :numeric
##
##
##
## start_station_name start_station_id end_station_name end_station_id
## Length:5803720 Length:5803720 Length:5803720 Length:5803720
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## start_lat start_lng end_lat end_lng
## Min. :41.64 Min. :-87.84 Min. : 0.00 Min. :-88.14
## 1st Qu.:41.88 1st Qu.:-87.66 1st Qu.:41.88 1st Qu.:-87.66
## Median :41.90 Median :-87.64 Median :41.90 Median :-87.64
## Mean :41.90 Mean :-87.65 Mean :41.90 Mean :-87.65
## 3rd Qu.:41.93 3rd Qu.:-87.63 3rd Qu.:41.93 3rd Qu.:-87.63
## Max. :42.07 Max. :-87.52 Max. :42.37 Max. : 0.00
## NA's :5855 NA's :5855
## member_casual ride_length day_of_week
## Length:5803720 Length:5803720 Min. :1.000
## Class :character Class1:hms 1st Qu.:2.000
## Mode :character Class2:difftime Median :4.000
## Mode :numeric Mean :4.103
## 3rd Qu.:6.000
## Max. :7.000
## cyclistic_one_yr_data_3$month <- month(mdy(cyclistic_one_yr_data_2$start_date))
cyclistic_one_yr_data_3 <- cyclistic_one_yr_data_3 %>%
mutate('month_name' = month.abb[month])
cyclistic_one_yr_data_3$start_date_dow <- mdy(cyclistic_one_yr_data_3$start_date)
cyclistic_one_yr_data_3$start_date_dow <- weekdays(cyclistic_one_yr_data_3$start_date_dow) - Created a tibble to look at the total number of rides based off day of the week.
- Created supplemetnal columns regarding date, month, day, year, day_of_week.
- Additionally, reformatted the started_at and ended_at columns as POSIXct columns in order to use both create a new column (ride_duration), which calculates the length of the ride.
- Lastly, created a tibble (bike_rider_type) to verify that there are two types of riders (casual and member) and three types of bikes (classical, docked, electric)
#Sys.setlocale("LC_TIME", "")
day_of_week_totalrides <- cyclistic_one_yr_data_3 %>%
group_by(start_date_dow) %>%
dplyr::summarize(total = n()) %>%
arrange(ordered(start_date_dow, levels = c("Sunday","Monday","Tuesday","Wednesday",
"Thursday","Friday","Saturday")))
cyclistic_one_yr_data_3$date <- as.Date(cyclistic_one_yr_data_3$start_date, "%m/%d/%y")
cyclistic_one_yr_data_3$Month <- format(as.Date(cyclistic_one_yr_data_3$date), "%m")
cyclistic_one_yr_data_3$day <- format(as.Date(cyclistic_one_yr_data_3$date), "%d")
cyclistic_one_yr_data_3$year <- format(as.Date(cyclistic_one_yr_data_3$date), "%y")
cyclistic_one_yr_data_3$Day_Of_Week <- format(as.Date(cyclistic_one_yr_data_3$date), "%A")
started_at <- cyclistic_one_yr_data_3$started_at
cyclistic_one_yr_data_3$started_at <- as.POSIXct(paste0(started_at, ':00'), format="%m/%d/%y %H:%M:%S")
ended_at <- cyclistic_one_yr_data_3$ended_at
cyclistic_one_yr_data_3$ended_at <- as.POSIXct(paste0(ended_at, ':00'), format="%m/%d/%y %H:%M:%S")
cyclistic_one_yr_data_3$ride_duration <- difftime(cyclistic_one_yr_data_3$ended_at,
cyclistic_one_yr_data_3$started_at)
bike_rider_type <- cyclistic_one_yr_data_3 %>%
dplyr::group_by(rideable_type, member_casual)
dplyr::summarize(bike_rider_type)## `summarise()` has grouped output by 'rideable_type'. You can override using the
## `.groups` argument.## # A tibble: 5 × 2
## # Groups: rideable_type [3]
## rideable_type member_casual
## <chr> <chr>
## 1 classic_bike casual
## 2 classic_bike member
## 3 docked_bike casual
## 4 electric_bike casual
## 5 electric_bike member- docked_bike refers to when bikes were taken out of docks and QC check was performed.Therefore, docked_bike isn’t pertinent to the study, so it needs to be removed from cyclistic_one_yr_data_3
- I need to cross reference the cumulative number of docked_bike so that I verify the correct amount is removed. I did this by filtering for when rideable_type is not equal to docked_bike.
is.factor(cyclistic_one_yr_data_3$ride_duration)## [1] FALSEcyclistic_one_yr_data_3$ride_duration <- as.numeric(as.character(cyclistic_one_yr_data_3$ride_duration))
is.numeric(cyclistic_one_yr_data_3$ride_duration)## [1] TRUEcyclistic_one_yr_data_4 <- cyclistic_one_yr_data_3 %>%
filter(rideable_type != "docked_bike")- Found NA values and then removed start_station_name, end_station_name, start_station_id, end_station_id columns because latitude and longitude columns can be used to determine ride paths. Doing this will also make the dataset smaller, so that I can use Tableau public ideally to make more thorough visualizations.
sum(is.na(cyclistic_one_yr_data_4$start_station_name))## [1] 839082sum(is.na(cyclistic_one_yr_data_4$end_station_name))## [1] 893698cyclistic_one_yr_data_5 <- cyclistic_one_yr_data_4- Convert ride_duration_hms from time object to double in order to count how many NA values are present.
sum(is.na(cyclistic_one_yr_data_5$start_station_id))## [1] 839214sum(is.na(cyclistic_one_yr_data_5$end_station_id))## [1] 893839cyclistic_one_yr_data_6 <- cyclistic_one_yr_data_5 %>%
select(-c(start_station_id, end_station_id))
cyclistic_one_yr_data_6b <- cyclistic_one_yr_data_6 %>%
mutate(ride_duration_hms = hms::hms(seconds_to_period(ride_duration)))
cyclistic_one_yr_data_6b$ride_duration_hms <- seconds(hms(cyclistic_one_yr_data_6b$ride_duration_hms))## Warning in .parse_hms(..., order = "HMS", quiet = quiet): Some strings failed to
## parse, or all strings are NAscyclistic_one_yr_data_6b$ride_duration_hms <- period_to_seconds(cyclistic_one_yr_data_6b$ride_duration_hms)
sum(is.na(cyclistic_one_yr_data_6b$ride_duration_hms))## [1] 88- Dropped NA values from ride_duration_hms and saved as a new data frame.
cyclistic_one_yr_data_6b_no_na <- cyclistic_one_yr_data_6b[!is.na(cyclistic_one_yr_data_6b$ride_duration_hms),]
cyclistic_one_yr_data_6b_no_na <- cyclistic_one_yr_data_6b_no_na[cyclistic_one_yr_data_6b_no_na$ride_duration_hms >= 0, ]- Find the minimum, maximum, median, and mean to ensure there are no negative ride_duration_hms values.
min(cyclistic_one_yr_data_6b_no_na$ride_duration_hms)## [1] 0max(cyclistic_one_yr_data_6b_no_na$ride_duration_hms)## [1] 90060median(cyclistic_one_yr_data_6b_no_na$ride_duration_hms)## [1] 598mean(cyclistic_one_yr_data_6b_no_na$ride_duration_hms)## [1] 944.7706descriptive_analysis_summary <- summary(na.omit(cyclistic_one_yr_data_6b_no_na$ride_duration_hms))
descriptive_analysis_summary## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0 338.0 598.0 944.8 1034.0 90060.0descriptive_analysis_summary_whole <- lapply(descriptive_analysis_summary, as.integer)- Gather data and arrange it into a data frame all by type of
rider.
- Minimum ride_duration_hms
- Maximum ride_duration_hms
- Median ride_duration_hms
- Mean ride_duration_hms
- Average ride_duration_hms per Day_Of_Week
- Gather data and arrange it into a data frame all by type of
rider.
ride_duration_hms_min <- aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration_hms ~ cyclistic_one_yr_data_6b_no_na$member_casual, FUN = "min")
ride_duration_hms_max <- aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration_hms ~ cyclistic_one_yr_data_6b_no_na$member_casual, FUN = "max")
ride_duration_hms_median <- aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration_hms ~ cyclistic_one_yr_data_6b_no_na$member_casual, FUN = "median")
ride_duration_hms_mean <- aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration_hms ~ cyclistic_one_yr_data_6b_no_na$member_casual, FUN = "mean")avg_ride_time_day_user <- aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration_hms ~ cyclistic_one_yr_data_6b_no_na$member_casual + cyclistic_one_yr_data_6b_no_na$Day_Of_Week, FUN = "mean")
aggregated_data <- data.frame(ride_duration_hms_min, ride_duration_hms_max, ride_duration_hms_median,
ride_duration_hms_mean, avg_ride_time_day_user)- Ordered Day_Of_Week to prepare data for visualization.
cyclistic_one_yr_data_6b_no_na$Day_Of_Week <- ordered(cyclistic_one_yr_data_6b_no_na$Day_Of_Week,
levels = c("Sunday", "Monday", "Tuesday",
"Wednesday", "Thursday",
"Friday", "Saturday"))
avg_ride_time_day_user_ordered_Day_Of_Week <-aggregate(cyclistic_one_yr_data_6b_no_na$ride_duration ~ cyclistic_one_yr_data_6b_no_na$member_casual + cyclistic_one_yr_data_6b_no_na$Day_Of_Week, FUN = "mean")- Created new data frame (cyclistic_one_yr_data_organized) for the purpose of grouping by rider type and the day of the week. Added a new column number_of_rides and average_duration
cyclistic_one_yr_data_organized <- cyclistic_one_yr_data_6b_no_na %>%
group_by(member_casual,Day_Of_Week) %>%
dplyr::summarize(number_of_rides = n(), average_duration = mean(ride_duration_hms)) %>%
arrange(member_casual, Day_Of_Week)## `summarise()` has grouped output by 'member_casual'. You can override using the
## `.groups` argument.cyclistic_one_yr_data_organized_2 <- cyclistic_one_yr_data_organized %>%
mutate(average_duration_min = average_duration/60)
cyclistic_one_yr_data_6b_no_na_fin <- cyclistic_one_yr_data_6b_no_na %>%
mutate(average_duration_min = ride_duration/60)- Created a plot of the new data frame, which shows the number of rides by type of rider and arranged by day of the week.
plot_total_rides <- ggplot(cyclistic_one_yr_data_organized_2, aes(x = Day_Of_Week, y = number_of_rides, fill = member_casual)) + geom_col(position = "dodge") + labs(title = "Total Number of Rides by Type of Rider", x = "Day of the Week", y = "Total Number of Rides", fill = "Type of Rider") + theme(plot.title = element_text(hjust = 0.5))
plot_total_rides
* + Created another plot showing the average ride duration by type of
rider and arranged by day of the week.
plot_avg_ride_duration<- ggplot(cyclistic_one_yr_data_organized_2, aes(x = Day_Of_Week, y = average_duration_min, fill = member_casual)) + geom_col(position = "dodge") + labs(title = "Average Ride Duration by Type of Rider", x = "Day of the Week", y = "Average Ride Duration (min)", fill = "Type of Rider") + theme(plot.title = element_text(hjust = 0.5))
plot_avg_ride_duration
#```{r}
write.csv(cyclistic_one_yr_data_6b_no_na_fin, “/Users/bethanyleach/Desktop/12_months_cycle_data_organized_final_2.csv”, row.names=FALSE)
#{r setup, include=FALSE} knitr::opts_chunk$set(echo=TRUE) library(knitr) #
Step 5 - Share
- I exported the cleaned dataset into a .csv file for further analysis in Tableau.
Bike Passengers (General)
- In this dataset, there were more members than casual bike users (~3.5 vs ~2.1 million).
Bike Passengers (Daily)
- Nearly twice as many members than casual users took bike rides during the week (Monday-Friday) whereas the delta between the two groups of users was much smaller during the weekend. This implies that members use the bikes to commute during the work week whereas casual users prefer to take rides recreationally on the weekend.
Bike Passengers (Monthly)
- On a monthly basis, members took more rides than casual users. There was a huge uptick in rides taken regardless of user type in the Spring/Summer, which makes sense due to the warmer temperatures.
Average Ride Duration
- On average, casual users’ rides were 8 mins longer than those of members’. Casual users always had longer rides than members regardless of the day of the week.
Bike Type Breakdown
- With respect to number of rides taken, roughly 1.5 times the number of casual users preferred electric over casual bikes. However, casual users took roughly 1.5 times longer rides on classic bikes than they did on electric bikes.
- Members took almost the same number of rides on electric bikes as they did on classic models. However, members took roughly 1.25 times longer rides on classic bikes than they did on electric bikes. + Overall, both demographics of users preferred to (on average) take longer rides on classic bikes.
Ride Start Time (Weekday)
- There was a huge uptick in total number of rides taken by members from 7-9am and 3-7pm. This makes sense since these times correlate with both the morning and evening work commute.
- There was a gradual increase in total number of rides taken by casual users throughout the day with the highest number of rides being taken at 5pm.
Ride Start Time (Weekend)
- The graph for the start time for both groups of riders was much more sinusoidal on the weekend with the valley/low point occuring in the morning at 5am and the peak/high point occuring in the afternoon at 3pm.
Step 6 - Act
Conclusion
- One key takeaway is that casual users take more rides than members during the weekend and they also take longer rides (on average). This implies that casual users are taking rides for leisure rather than for commutes.
- There is an uptick is total number of rides during the morning and evening commute for both demographics of users. The difference is the commute spikes for members are more drastic than those for casual users.
- Both members and casual users alike take more rides in Spring and Summer.
- Casual users on average take longer rides on classic bikes than they do electric bikes, but casual users took more individual rides on electric bikes than they did on classic bikes.
Recommendation for Cyclistic’s marketing team
- I recommend a digital marketing campaign (app, email) targeted to casual riders for a limited period of time offering them discounts once they opt for an annual membership. In order to fixate on the busiest days for casual users, Cyclistic should release discounts for this campaign on the weekend (Friday, Saturday, Sunday).
- In addition, Cyclisitic could try distributing buddy passes on the weekend to casual users in order to reach a wider audience and possibly obtain more users.
- The map dashboard shows that the majority of both casual users and members rent bikes in downtown Chicago near parks. Therefore, I would suggest that Cyclistic focus on public parks and fitness centers when launching this marketing campaign. Traveling via car, public transportation, and by foot is very time consuming, but bikes are both environmentally conscious and efficient.
- I would also suggest that Cyclistic analyze their current pricing to determine which if an annual membership is actually more cost-effective and if it’s not, I would recommend lowering the price in order to procure more members from the casual cohort.
- Lastly, I would suggest creating a seasonal membership. Chicago is very cold in the winter, but with a seasonal memebership, Cyclistic would be able to reach a wider audience. They may be able to obtain more members since people wouldn’t feel like they’re wasting an annual membership for half of the year.