Cyclistic - Bike Share
Tom Armstrong
2022-10-15
Prelude
The following report is an analysis of data for a fictional company. It is the capstone project for the “Google Data Analytics” course offered through corsera and my first attempt at a project liek this with Rstudio. I would appreciate any thoughts, opinions and feedback on any process i have used throughout.
Scenaro
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.
Deliverables
- A clear statement of the business task
- A description of all data sources used
- Documentation of any cleaning or manipulation of data
- A summary of your analysis
- Supporting visualizations and key findings
- Your top three recommendations based on your analysis
ASK
What is the problem ?
The team at cyclistic understand that the annual membership clients are more profitable than the casual clients. Is there a way we can encourage more casual clients to purchase annual subscriptions ? We need to understand the differences between how both the casual and membership clients use the cyclistic service. This knowledge will provide the foundation for a more effective marketing campaign to encourage casual clients to make the switch.
The Business Task
How do annual members and casual riders use Cyclistic bikes differently?
PREPARE
The Data
The data is information pertaining to the use of Cyclistics services between the time period of 1/8/21 to 31/07/22. The data is in the format of .CSV files in which there is a file for each month, 12 in total.
Lisencing
The data for this analysis has been collected autonomously and provided by Cyclistic under license linked below. https://ride.divvybikes.com/data-license-agreement
Credibility
The data has been generated autonomously by the riders as they use the service. Information such as start, end time, bike type, client type and location are logged. With this being said, bias will not be an issue throughout this analysis. The data is a true representation of the activity throughout the time period.
Privavcy and Security
I understand the responsibilities of a data analysis in regards to data privacy and security. This work will not been viewed by any other individuals other than those approved by key stakeholders within the company.
How will it help ?
I am confident a wealth of information can be extracted from the data to better understand how the usage of client type differs. There are no problems with the data and the integrity apears solid.
PROCESS
Tools
I have chosen to use R for my cleaning organizing and transforming. I am most comfortable with this tool and feel it provides me with the most effective platform to get the crux of the work done quickly.
Organizing, Cleaning and Transforming Log.
These are the steps i took to ensure the data is organised, clean and transformed, makeing sure i can extract as much of the information from it as possible.
File Organisation
- Import files (12)
- Create list of files
- Bind 12 files into one data frame (df)
library(tidyverse)
library(janitor)
library(lubridate)
library(scales)
library(ggthemes)
library(RColorBrewer)
library(viridis)
library(rmarkdown)
library(ggmap)
library(plotly)
library(stringr)
library(shadowtext)
library(gt)dfv7 <- read.csv("cyclistic_clean_v5")
file.names <- list.files(pattern = "*.csv")
df_lists <- lapply(file.names,read.csv)
df <- bind_rows(df_lists)
Cleaning
- Remove unwanted columns within combined data frame.
- Removed columns were missing information and decided they were not required for this analysis. These included “start” and “End” location name. We still have a lat. and long. of these locations so we can run some location base analysis using these variables.
- Checked for duplicate entries using the “ride_id” column.
- Removed 226723 rows that were associated with a “docked bike” within the “rideable_type variable.
df <- df %>%
select(-c(start_station_id,start_station_name,
end_station_id,end_station_name))
n_distinct(df$ride_id)
dfv5 <- dfv4[!(dfv4$rideable_type=="docked_bike"),]
Transformation
Created some new columns from the date/time column to increase the detail in which we can interrogate the data.
These include; Date,Day,Month,Year,Month and Year and Weekday.
Changed datetime column from a chr to a datetime variable so i can use the difftime function to find the ride_time in mins.
Changed the day and month columns to num variables to assist in analysis.
Changed difftime varibale that was associated with the ride_time column to a num variable.
Ride_time column has negative values (149)
The top 1% of the values in ride_time are above a one hour.
I have decided to just remove the negative values and explore the high values later.
Decided to remove top and bottom 1 % to use 98 % of data. This removes some of the more dramatic outliers.
Check for values < 0
Remove values <0
Check percentiles for outliers
Remove top an bottom 1%
Removed 114579 rows as outliers.”
re-ordering days of week within in the weekday column
Rounding lat and lng columns to assist in cleaning up the ploting ability.
df$date <- as.Date(df$started_at)
df$day <- format(as.Date(df$date),"%d")
df$month <- format(as.Date(df$date),"%m")
df$year <- format(as.Date(df$date),"%y")
df$month_year <- format(as.Date(df$date),"%h - %y")
df$weekday <- format(as.Date(df$date),"%A")
df$started_at_ <- ymd_hms(df$started_at)
df$ended_at_ <- ymd_hms(df$ended_at)
df$ride_time <- difftime(df$ended_at_,df$started_at_, unit = "mins")
df$dayI <- as.numeric(df$day)
df$monthI <- as.numeric(df$month)
is.factor(df$ride_time)
df$ride_time <- as.numeric(as.character(df$ride_time))
as.numeric(df$ride_time)
nrow(df[df$ride_time <0, ])
dfv3 <- df[!(df$ride_time<0),]
ventiles = quantile(df$ride_time, seq(0, 1, by=0.01))
ventiles
dfv7 <- dfv5 %>%
filter(ride_time > as.numeric(ventiles['1%'])) %>%
filter(ride_time < as.numeric(ventiles['99%']))
df$weekday <-ordered(df$weekday,levels=c("Monday","Tuesday","Wednesday",
"Thursday","Friday","Saturday",
"Sunday"))
random_sample$start_lat_r <- round(random_sample$start_lat, digits = 3)
random_sample$start_lon_r <- round(random_sample$start_lng, digits = 3)
Analyse and Share
Key take aways (For those in a hurry !)
- Time period - Aug 21 to Jul 22.
- 5.5 M individual rides.
- 40 % of those clients in were casual clients.
- Casuals most likely to ride on weekend between 3 pm and 6 pm.
- Casuals ride for the longest on average.
- Outside of the CBD, the hots-spots for casual clients to start their ride is Near south side, West town and Lincoln park.
Distribution of Client types by time
dfv7 %>%
group_by(member_casual) %>%
summarise(count = length(ride_id),
pct = length(ride_id)/nrow(dfv7)) %>%
ggplot(aes(x=member_casual, y =count, fill = member_casual))+
geom_col()+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
labs(title = "Proportion of client type",
subtitle = "Time period (Aug 21 - Jul 22)",
x = "Client Type",
y = "Count") +
scale_y_continuous(labels = comma)+
theme_economist()+
theme(legend.position = "none",
axis.title.y = element_text(vjust = 3),
axis.title = element_text(face = "bold", size = 12),
axis.text = element_text(face = "bold", size = 12),
title = element_text(face = "bold",size = 14),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
annotate(geom = "text", x= 1.00, y =1200000, label = "2.2 M\n 40% ",
color = "white", size = 15, fontface = "bold")+
annotate(geom = "text", x = 2.03, y= 1700000, label = "3.3 M \n 60 %",
color = "white", size = 15, fontface = "bold")+
scale_x_discrete(labels=c("casual" = "Casual", "member" = "Member"))
table_1 <- dfv7 %>%
group_by(member_casual) %>%
summarise(Count = length(ride_id),Pct = Count/nrow(dfv7)*100) %>%
adorn_totals(where = "row", name = "Totals")
table_1 <- table_1 %>%
gt() %>%
tab_header(title = "Proportions of client type") %>%
cols_label(member_casual = "Client", Count = "Count", Pct = "%") %>%
tab_style(style = list(cell_fill(color = "#F4F4F4")),
locations = cells_body(columns = member_casual))
table_1| Proportions of client type | ||
| Client | Count | % |
|---|---|---|
| casual | 2225158 | 40.02074 |
| member | 3334854 | 59.97926 |
| Totals | 5560012 | 100.00000 |
From a total of around 5.5 M rides over the past 12 months, this is the client type split for the period Aug 21 to Jul 22. Casual clients were at 40 % and Member clients 60 %.
#Reorder months
dfv7$month_year <- format(as.Date(dfv7$date),"%h-%y")
dfv7$month_year <- ordered(dfv7$month_year,levels=c ("Aug-21",
"Sep-21",
"Oct-21",
"Nov-21",
"Dec-21",
"Jan-22",
"Feb-22",
"Mar-22",
"Apr-22",
"May-22",
"Jun-22",
"Jul-22" ))
dfv7 <- dfv7 %>%
mutate(month_year_int = as.numeric(month_year))
dfv7 %>%
group_by(member_casual,month_year)%>%
summarise(count_month_year = length(ride_id),
PCT = count_month_year/nrow(dfv7)*100) %>%
ggplot(aes(x = month_year, y = count_month_year,
fill = member_casual, width = .9))+
geom_col(position = "dodge" )+
theme_economist()+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
labs(title = "Distribution of client type by month",
subtitle = "Time period (Aug 21 - Jul 22)",
x = " Month",
y = " Count")+
theme(legend.title = element_blank(),
axis.title = element_text(face = "bold", size = 11),
axis.text = element_text(face = "bold"),
axis.text.y = element_text(vjust = .2, hjust = 1),
axis.text.x = element_text(vjust = 2, hjust = .5),
title = element_text(face = "bold", size = 12),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"),
axis.title.x = element_text(vjust = -2),
axis.title.y = element_text(vjust = 3))+
scale_y_continuous(labels = comma)
casual_month <- dfv7%>%
group_by(month_year,member_casual)%>%
summarise(Count = length(ride_id),
PCT = Count/nrow(dfv7)*100) %>%
subset(member_casual=="casual")
member_month <- dfv7 %>%
group_by(month_year,member_casual)%>%
summarise(Count = length(ride_id),
PCT = Count/nrow(dfv7)*100) %>%
subset(member_casual=="member")
member_month <- member_month %>%
rename(count_member= Count, PCT_member = PCT, month = month_year) %>%
select(-c(member_casual))
casual_month<- casual_month %>%
rename(count_casual=Count, PCT_casual = PCT, month = month_year) %>%
select(-c(member_casual))
clients_month <- full_join(member_month,casual_month)
clients_month <- clients_month %>%
adorn_totals(where = "row", name = "Totals")%>%
gt()%>%
tab_header(title = "Distribution of client type by the month") %>%
cols_label(month = "Month",count_casual = "Casual", PCT_casual = "Casual %",
count_member = "Member", PCT_member= " Member %")%>%
tab_style(
style = list(cell_fill(color = "#F4F4F4")),
locations = cells_body(columns = month )
)
clients_month| Distribution of client type by the month | ||||
| Month | Member | Member % | Casual | Casual % |
|---|---|---|---|---|
| Aug-21 | 386882 | 6.958294 | 355522 | 6.3942668 |
| Sep-21 | 387479 | 6.969032 | 318686 | 5.7317502 |
| Oct-21 | 369507 | 6.645795 | 228278 | 4.1057106 |
| Nov-21 | 250175 | 4.499541 | 97175 | 1.7477480 |
| Dec-21 | 175778 | 3.161468 | 63346 | 1.1393141 |
| Jan-22 | 84035 | 1.511418 | 16983 | 0.3054490 |
| Feb-22 | 92624 | 1.665896 | 19314 | 0.3473734 |
| Mar-22 | 191643 | 3.446809 | 78927 | 1.4195473 |
| Apr-22 | 241330 | 4.340458 | 110603 | 1.9892583 |
| May-22 | 349650 | 6.288655 | 245153 | 4.4092171 |
| Jun-22 | 394367 | 7.092916 | 327750 | 5.8947715 |
| Jul-22 | 411384 | 7.398977 | 363421 | 6.5363348 |
| Totals | 3334854 | 59.979259 | 2225158 | 40.0207410 |
Looking at plot 2 “Distribution of client type by month”, we can see there is a predictable decline throughout the winter months for both client types. We can observe the usage of members remain proportionally higher than that of the casuals. We could hypothesize that this is the result of the members clients using the bikes to get to work.
#Reordering days of week
dfv7$weekday <-ordered(dfv7$weekday,levels=c("Monday","Tuesday","Wednesday",
"Thursday","Friday","Saturday",
"Sunday"))
plot2 <-dfv7 %>%
group_by(weekday,member_casual)%>%
summarise(count=length(ride_id),
pct = (count/nrow(dfv7)*100))
plot2$pct <- paste(round(plot2$pct,0),"%")
# dfv7 %>%
# group_by(member_casual,weekday)%>%
# summarise(count=length(ride_id),
# pct = count/nrow(dfv7)*100)
#pct = count/nrow(dfv7)*100)
plot2 %>%
ggplot((aes(x=weekday,y=count, fill= member_casual, width = .9)))+
geom_bar(stat="identity", position = "dodge")+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
labs(title = "Distribution of client type per day of week",
subtitle = "Time period (Aug 21 - Jul 22)",
x = "Day of Week",
y = " Count")+
theme_economist()+
theme(legend.title = element_blank(),
axis.title.x = element_text(vjust = -2),
axis.title = element_text(face = "bold", size = 11),
axis.title.y = element_text(vjust = 3),
title = element_text(face = "bold",size = 12),
axis.text = element_text(face = "bold", size = 10),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
scale_y_continuous(labels = comma)+
geom_text(aes(label = plot2$pct), position = position_dodge(0.8),hjust =.4, vjust = -0.2,
size = 4, color = "black", fontface= "bold")
dfv7$weekday <-ordered(dfv7$weekday,levels=c("Monday","Tuesday","Wednesday",
"Thursday","Friday","Saturday",
"Sunday"))
Casualv2 <- dfv7 %>%
group_by(weekday,member_casual) %>%
summarise(Count_casual = length(ride_id),
PCT_casual = Count_casual/nrow(dfv7)*100)%>%
subset(member_casual == "casual")
Casualv2 <- Casualv2 %>%
select(-c(member_casual))
Memberv2 <- dfv7 %>%
group_by(weekday,member_casual) %>%
summarise(Count_member = length(ride_id),
PCT_member = Count_member/nrow(dfv7)*100) %>%
subset(member_casual == "member")
Memberv2 <- Memberv2 %>%
select(-c(member_casual))
weekday_clients <- full_join(Casualv2,Memberv2)
weekday_clients<- weekday_clients %>%
adorn_totals(where = "row", name = "Totals") %>%
gt()%>%
tab_header(title = "Distribution of client type per day of week") %>%
cols_label(weekday = "Weekday",Count_casual = "Casual", PCT_casual = "Casual %",
Count_member = "Member", PCT_member= " Member %")%>%
tab_style(
style = list(cell_fill(color = "#F4F4F4")),
locations = cells_body(columns = weekday )
)
weekday_clients| Distribution of client type per day of week | ||||
| Weekday | Casual | Casual % | Member | Member % |
|---|---|---|---|---|
| Monday | 263032 | 4.730781 | 466454 | 8.389442 |
| Tuesday | 246600 | 4.435242 | 517047 | 9.299386 |
| Wednesday | 255259 | 4.590979 | 516049 | 9.281437 |
| Thursday | 285665 | 5.137849 | 516334 | 9.286563 |
| Friday | 310285 | 5.580653 | 460561 | 8.283453 |
| Saturday | 456117 | 8.203525 | 446757 | 8.035180 |
| Sunday | 408200 | 7.341711 | 411652 | 7.403797 |
| Totals | 2225158 | 40.020741 | 3334854 | 59.979259 |
Plot 3, “Distribution of client type per day of week” shows distribution of both client types throughout the week. We can observe relative stability of usage from the member clients on weekdays and a drop in usage on the weekend. Causal clients show a trend of increased usage though the week, peaking on the weekend. This might further suggest that the member clients are using the service to commute to and from work.
dfv7 %>%
group_by(member_casual, start_hour) %>%
summarise(count_hour = length(ride_id)) %>%
ggplot(aes(x=start_hour, y=count_hour, color = member_casual))+
geom_line( size = 3.5, alpha =.85) +
theme_economist()+
scale_color_manual(values = c("#B67B49", "#4984B6"))+
labs(title = "Distribution of clients by start hour",
subtitle = "Time period (Aug 21 - Jul 22)",
x = "Start Hour (24H Time)",
y = "Count")+
theme(legend.title = element_blank(),
axis.title = element_text(face = "bold"),
axis.text = element_text(face = "bold", size =13),
axis.text.x = element_text(size=13, vjust = 2),
axis.text.y = element_text(hjust =1.2, vjust = -.1),
title = element_text(face = "bold", size = 14),
axis.title.y = element_text(vjust = 3, size = 14),
axis.title.x = element_text(vjust = -2, size = 14),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
scale_y_continuous(labels = comma) 
Plot 4 is the “Distribution of both client types by start hour”. We can see the highest period of activity for both member and casual clients is between 3pm - 8pm. Member clients have a noticeable increase relative to the casual clients at the time periods of 6am - 9am and 3pm-8pm.
Distribution of Client types by bike type
dfv7 %>%
count(type = factor(rideable_type), client_type = factor(member_casual)) %>%
mutate(pct = prop.table(n)) %>%
ggplot(aes(x = type, y=pct, fill= client_type, label = scales::percent(pct))) +
geom_col (position = 'dodge')+
theme_economist()+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
geom_text(position = position_dodge(width = .9), # move to center of bars
vjust = -0.2, # nudge above top of bar
size = 4.5, fontface = "bold") +
scale_y_continuous(labels = scales::percent)+
labs(title = "Distribution of client by bike type",
subtitle = "Time period (Aug 21 - Jul 22)",
x = "Bike Type",
y= "")+
theme(legend.title = element_blank(),
axis.title = element_text(face = "bold", size = 11,vjust = .5),
title = element_text(face = "bold", size = 12),
axis.text = element_text(face = "bold"),
axis.text.x = element_text(size = 10),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
scale_x_discrete(labels = c("classic_bike" = " Classic Bike", "electric_bike" = "Electric Bike"))
dist_bike <- dfv7 %>%
count(Bike_type = factor(rideable_type),Client_type = factor(member_casual)) %>%
mutate("%"=prop.table(n)*100)
dist_bike <- dist_bike %>%
gt()%>%
tab_header(title = "Distribution of client by bike type") %>%
cols_label(Bike_type = "Bike Type",Client_type = "Client Type")%>%
tab_style(
style = list(cell_fill(color = "#F4F4F4")),
locations = cells_body(columns = Bike_type))#
dist_bike| Distribution of client by bike type | |||
| Bike Type | Client Type | n | % |
|---|---|---|---|
| classic_bike | casual | 1090593 | 19.61494 |
| classic_bike | member | 1900839 | 34.18768 |
| electric_bike | casual | 1134565 | 20.40580 |
| electric_bike | member | 1434015 | 25.79158 |
Plot 5 outlines the percentage “Distribution of client types by bike types.”
Average ride times
average_ride_times_biketype <- dfv7 %>%
group_by(member_casual,rideable_type)%>%
summarize(ave_type=mean(ride_time))
average_ride_times_biketype$ave_type <- round(average_ride_times_biketype$ave_type, digits = 2)
ggplot(average_ride_times_biketype,aes(x=rideable_type,y = ave_type,
fill=member_casual))+
theme_economist()+
geom_col(position = "dodge")+
geom_text(aes(x = rideable_type, y = ave_type, label = ave_type, size = 14),
stat = "sum", position = position_dodge(.9),vjust=-.2, show.legend = F)+
labs(title = "Average ride time for client and bike types",
subtitle = "Time period (Aug 21 - Jul 22)",
x = "Bike Type",
y = "Average ride minutes")+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
theme(legend.title = element_blank(),
axis.title = element_text(face = "bold", size = 10),
title = element_text(face = "bold", size = 12),
axis.title.y = element_text(vjust = 3, size = 12, face = "bold"),
axis.title.x = element_text(size = 12),
axis.text.x = element_text(size = 10.5, face = "bold"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
scale_x_discrete(labels = c("classic_bike" = " Classic Bike", "electric_bike" = "Electric Bike"))
#table showing average ride times.Plot 6 shows the average ride time for each client and bike type. We can see a higher average ride time for casual clients compared with member clients. Lets explore how this trend plays out by the month and weekday .
dfv7$month_year <- format(as.Date(dfv7$date),"%h-%y")
dfv7$month_year <- ordered(dfv7$month_year,levels=c ("Jul-22",
"Jun-22",
"May-22",
"Apr-22",
"Mar-22",
"Feb-22",
"Jan-22",
"Dec-21",
"Nov-21",
"Oct-21",
"Sep-21",
"Aug-21" ))
month_year_ride <- dfv7 %>%
group_by(month_year, member_casual, rideable_type) %>%
summarize(count = n(),
ave = mean(ride_time))
ggplot(month_year_ride, mapping = aes(x = month_year, y = ave,
fill = member_casual))+
geom_col(position = "dodge")+
theme_economist()+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
labs(title = "classic")+
facet_wrap(~rideable_type)+
labs(title = "Average ride time by month for client and bike types",
subtitle = "Time period (Aug 21 - Jul - 22)",
y = "Average ride minutes",
x = "Month")+
theme(legend.title = element_blank(),
axis.title.y = element_text(vjust = 3,size = 13),
axis.title.x = element_text(vjust = -1),
title = element_text(face = "bold", size = 13),
axis.text = element_text(size = 12, face = "bold"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
coord_flip()
Plot 7 shows the average ride time in minutes for each month for both client types. The chart on the left represents the classic bike, on the right the electric bike.
#col
dfv7$weekday <-ordered(dfv7$weekday,levels=c("Sunday","Saturday","Friday",
"Thursday","Wednesday","Tuesday",
"Monday"))
dfv7 %>%
group_by(weekday,rideable_type,member_casual) %>%
summarize(ave_ride_week = mean(ride_time),
count_ = length(ride_id)) %>%
ggplot(mapping = aes(x = weekday, y = ave_ride_week,fill = member_casual))+
geom_col(position = "dodge")+
facet_wrap(~rideable_type)+
labs(title = "Average ride time by weekday for client and bike type",
subtitle = "Time period (Aug 21 - Jul 22)",
x="Weekday",
y="Average ride minutes")+
theme_economist()+
scale_fill_manual(values = c("#B67B49", "#4984B6"))+
theme(legend.title = element_blank(),
axis.title = element_text(face = "bold", size = 12),
title = element_text(face = "bold", size = 13),
axis.title.x = element_text(vjust = -1),
axis.title.y = element_text(size =13),
axis.text.y = element_text(hjust = 1.2),
axis.text = element_text(face = "bold"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
panel.background = element_rect(fill = "whitesmoke"),
plot.background = element_rect(fill = "whitesmoke"))+
coord_flip()
Plot 8 shows the average ride time by weekday for both client and bike types. We can see the average ride time for member clients throughout the week is around 13 mins for both bike types, peaking slightly on the weekend. Casual clients have a higher overall ride time than that of members about 18 mins with average ride time on the classic bike`s being the longest.
Geo Location
library(sp)
library(leaflet)
#Random Sample
random_sample <- sample_n(dfv7,100000)
random_sample$start_lat_r <- round(random_sample$start_lat, digits = 3)
random_sample$start_lon_r <- round(random_sample$start_lng, digits = 3)
#Subset of each client tyoe
ran_cas <- random_sample %>%
filter(member_casual == "casual")
ran_mem <- random_sample %>%
filter(member_casual == "member")
factpal <- colorFactor(topo.colors(2), random_sample$rideable_type)
#Leaflet
m <- leaflet() %>%
setView(lat = 41.888,lng = -87.629, zoom=11)%>%
addTiles(group = "OSM (default)")%>%
addProviderTiles(providers$Stamen.TonerLite, group = "Light") %>%
addProviderTiles(providers$CartoDB.DarkMatter, group = "Dark") %>%
addCircles(data = ran_cas, lng = ~start_lon_r,
lat = ~start_lat_r,
popup = ~paste("<h3> Info:</h3>","<b>Bike Type:</b>",
rideable_type,"<br>","<b>Ride Time:</b>",
ride_time, sep = " "),color = "#B67B49",opacity = .01,
radius = 150, group = "Casual",fillOpacity = .01,stroke = F) %>%
addCircles(data = ran_mem, lng = ~start_lon_r,
lat = ~start_lat_r,
popup = ~paste("<h3> Info:</h3>","<b>Bike Type:</b>",
rideable_type,"<br>","<b>Ride Time:</b>",
ride_time, sep = " "),color = "#4984B6",
opacity = .01, radius = 150,group = "Member",fillOpacity = .01,
stroke = F) %>%
addCircleMarkers(data = ran_cas, lng = ~start_lon_r,
lat = ~start_lat_r,
popup = ~paste("<h3> Info:</h3>","<b>Bike Type:</b>",
rideable_type,"<br>","<b>Ride Time:</b>",
ride_time, sep = " "),
clusterOptions = markerClusterOptions(), group = "Casual_Cluster")%>%
addCircleMarkers(data = ran_mem, lng = ~start_lon_r,
lat = ~start_lat_r,
popup = ~paste("<h3> Info:</h3>","<b>Bike Type:</b>",
rideable_type,"<br>","<b>Ride Time:</b>",
ride_time, sep = " "),
clusterOptions = markerClusterOptions(), group = "Member_Cluster") %>%
addCircleMarkers(data = random_sample,lng = ~start_lon_r,lat=~start_lat_r,
color = ~factpal(rideable_type) ,
popup = ~paste("<h3> Info:</h3>","<b>Bike Type:</b>",
rideable_type,"<br>","<b>Ride Time:</b>",
ride_time, sep = " "),
group = "Bike Type",opacity = 0.01,fillOpacity = 0.01, stroke = F) %>%
addLegend(pal = factpal, position = "bottomright", values= random_sample$rideable_type, group = "Bike Type")%>%
addLayersControl(overlayGroups = c("Casual","Casual_Cluster", "Member","Member_Cluster","Bike Type"),
baseGroups = c("OSM (default)", "Light", "Dark"),
options=layersControlOptions(collapsed = F)) %>%
hideGroup(c("Member", "Member_Cluster", "Bike Type"))
#Call map
m