How to hard code the color for the displayed variable in plotly pie chart?

Question

Do we have something similar to our ggplot where we define the color panel for each variable ?

scale_manual <- function(...){
ggplot2::manual_scale(
"fill",
values = setnames(c("green","red","blue","yellow","grey"),
c("var1","var2","var3","var4","var5")),
...
)
}

Although this Q seems to answer, How can I change the colors of the slices in pie charts in plotly for r using hexadecimal strings?

but it is not working here.

Please consider an reprex below:

library(plotly)

USPersonalExpenditure <- data.frame("Categorie"=rownames(USPersonalExpenditure), USPersonalExpenditure)
data <- USPersonalExpenditure[,c('Categorie', 'X1960')]

p <- plot_ly(data, labels = ~Categorie, values = ~X1960, type = 'pie') %>%
  layout(title = 'United States Personal Expenditures by Categories in 1960',
         xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
         yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))

#trial 1
plot_ly(data, labels = ~Categorie, values = ~X1960, type = 'pie', marker = list(color = rainbow(5))) %>%
layout(title = 'United States Personal Expenditures by Categories in 1960', 
xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))

# trial 2
plot_ly(data, labels = ~Categorie, values = ~X1960, type = 'pie', marker = list(color = brewer_pal(5, "Set3"))) %>%
layout(title = 'United States Personal Expenditures by Categories in 1960', 
xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))

Since there are many plots using the same data, color need to be consistent.

So, trying to hard code for each variable.

maaniB · Accepted Answer · 2019-08-22T11:45:51.003

4

I usually use Color mapping functions from the leaflet package before making interactive plots for shiny apps and plotly charts:

Hard code the color variable using your desired palette.

data$color <- leaflet::colorFactor(
  palette = "Dark2", domain = data$Categorie
  )(data$Categorie)

plot_ly(
  data, labels = ~Categorie,  values = ~X1960, type = 'pie',
  marker = list( colors = ~color)
  ) %>%
  layout(
    title = 'United States Personal Expenditures by Categories in 1960', 
    xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
    yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE)
    )

You may also hardcode it manually:

colors_list <- list(
  "Food and Tobacco" = "#1B9E77",
  "Household Operation" = "#D95F02",
  "Medical and Health" = "#7570B3",
  "Personal Care" = "#E7298A",
  "Private Education" = "#66A61E"
)

data$color <- dplyr::recode(data$Categorie, !!!colors_list)


plot_ly(
  data, labels = ~Categorie,  values = ~X1960, type = 'pie',
  marker = list( colors = ~color)
) %>%
  layout(
    title = 'United States Personal Expenditures by Categories in 1960', 
    xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
    yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE)
  )

edited Aug 22 '19 at 11:45

answered Aug 22 '19 at 08:08

maaniB

595
6
23

Thanks for adding solution. currently, unable if install leaflet tried both install.packages("leaflet") and (!require('devtools')) install.packages('devtools') devtools::install_github('rstudio/leaflet') Issue with "png" package itself – Abhishek Aug 22 '19 at 11:33
@Abhi leaflet is a very useful and popular package. You can install its dependencies manually. What is your `Sys.info()`? Also, I'm gonna add another method for hardcoding the colors by editing my response. – maaniB Aug 22 '19 at 11:39
Error starts with PNG. read.c:3:17: fatal error: png.h: No such file or directory #include ^ compilation terminated. make: *** [read.o] Error 1 ERROR: compilation failed for package ‘png’ * removing ‘/home/myUserID/R/x86_64-redhat-linux-gnu-library/3.4/png’ Warning in install.packages : installation of package ‘png’ had non-zero exit status ERROR: dependency ‘png’ is not available for package ‘leaflet’ * removing ‘/home/myuserID/R/x86_64-redhat-linux-gnu-library/3.4/leaflet’ Warning in install.packages : installation of package ‘leaflet’ had non-zer – Abhishek Aug 22 '19 at 11:44
close your rstudio and do not save the workspace image. remove png package in your rstudio (`remove.packages("png", lib="~/home/myUserID/R/x86_64-redhat-linux-gnu-library/3.4")`) and delete it from your `/home/myUserID/R/x86_64-redhat-linux-gnu-library/3.4/png`. Then, `install.packages("png")`. If it is successful, you can install `leaflet` – maaniB Aug 22 '19 at 11:49
thanks for helping me.... It was an issue with png.... https://stackoverflow.com/questions/21800909/cannot-build-r-package-png-fedora-20 – Abhishek Aug 22 '19 at 11:50
If all the problems are solved and you think my solution is enough you can accept it :-) Good luck. – maaniB Aug 22 '19 at 11:52
How can we disable the legend ? describing the color ? – Abhishek Aug 26 '19 at 10:21

How to hard code the color for the displayed variable in plotly pie chart?

1 Answers1