CODEWITHSUNDEEP
rggplot2
HABLOH
HABLOH

Reputation: 470

R and ggplot: shade the color of the categorical variable

I would like to make a graph that represents the frequency of an event per hour of each day of the week. How do I change the color of the day of the week? I would like to switch from black to red to the proximity of the weekend

  weekday_hour_pickup <- my_data %>%
      mutate(hour_pick = hour(new_pickup))%>%
      group_by(hour_pick, weekday) %>%
      count() %>%
      ggplot(aes(hour_pick, n, color = weekday)) +
      geom_line(size = 1.5) +
      labs(x = "hour of the day", y = "count")

  plot(weekday_hour_pickup)

enter image description here

Upvotes: 3

Views: 1570

Answers (2)

StupidWolf
StupidWolf

Reputation: 47008

First, to simulate what your data might look like

TIMES = seq(from=mdy_hm("May 11, 1996 12:05"),
to=mdy_hm("May 18, 1996 12:05"),length.out=10000)

days.of.week <- weekdays(as.Date(4,"1970-01-01",tz="GMT")+0:6)
n = 1000
set.seed(100)
i = sample(1:length(TIMES),n)
my_data <- data.frame(
new_pickup=TIMES[i],
weekday = weekdays(TIMES[i])
)

Before you do plot, you need to factor the weekday variable in the correct order:

my_data$weekday <- factor(my_data$weekday,levels=days.of.week)

Then we set the color, in this case I set the extremes to be black and firebrick. You might needa play around with this

COLS <- alpha(colorRampPalette(c("black","firebrick"))(7),0.6)
names(COLS) <-  days.of.week

Then we plot

weekday_hour_pickup <- my_data %>%
      mutate(hour_pick = hour(new_pickup))%>%
      group_by(hour_pick, weekday) %>%
      count() %>%
      ggplot(aes(hour_pick, n, color = weekday)) +
      geom_line(size = 1) +
      labs(x = "hour of the day", y = "count")+
      scale_color_manual(values=COLS) + 
      theme_bw()

Not a very nice plot because I don't have your data, but you can adjust the COLS to get the contrast you need.

enter image description here

Upvotes: 1

Edo
Edo

Reputation: 7858

This a good option to make your descrite values as if they have a gradient

# call libraries
library(dplyr)
library(ggplot2)

# define weekdays
weekdays <- c("Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", "Sunday")

# define 7 colours from black to red
colours  <- scales::seq_gradient_pal(low = "#000000", high = "#ff0000", space = "Lab")(1:7/7)

# create values for scale colour manual
values <- setNames(colours, weekdays)

# create fake data
data <- tibble(hour_pick = rep(1:24, 7), 
               n         = rnorm(7*24) + rep(1:7, each = 24), 

               # weekday must be a factor if you want your legend to be sorted!
               weekday   = factor(rep(weekdays, each = 24), levels = weekdays, ordered = TRUE))

ggplot(data) +
  geom_line(aes(x = hour_pick, y = n, colour = weekday)) +
  theme_light() +
  scale_colour_manual(values = values)

enter image description here

Upvotes: 2

Related Questions