Modify color and font of emmip plot (emmeans package) in R

Question

Hello :) I am desperately trying to change the colors and font of my emmip plot (plot from the emmeans package in R) but none of my codes are working. Currently my code for the plot looks like this:

emmip(Model, group ~ gend, CIs=TRUE, nuisance = c("known", "age_dup", "edu"), 
      xlab = "", 
      ylab = "Intention to use the platform")

I read in the manueal from the R-package that the code from emmip() can be combined with ggplot2 codes. But when I add the following two codes (that I successfully use in another ggplot) - nothing changes in my plot:

+ theme(text=element_text(family="serif", size=13)
+ scale_fill_brewer(palette="Blues"))

I varied them already, for example "," instead of "+"

Does anyone have an idea how I can make these two modifications work in emmip? Thank you all in advance!

Here is the dput of my data (first 30 rows):

structure(list(dv = c(1, 5, 5, 1, 3, 5, 2, 1, 5, 5, 2, 4, 6, 
7, 3, 5, 5, 6, 7, 1, 7, 6, 2, 4, 7, 6, 5, 1, 6, 6), gend = structure(c(1L, 
2L, 1L, 1L, 1L, 2L, 1L, 1L, 2L, 1L, 1L, NA, 1L, 1L, 1L, 1L, 1L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 2L, 1L, 1L, 2L), .Label = c("Male", 
"Female"), class = "factor"), group = structure(c(5L, 3L, 5L, 
3L, 2L, 1L, 3L, 4L, 2L, 1L, 3L, 2L, 3L, 3L, 4L, 4L, 2L, 4L, 5L, 
5L, 1L, 4L, 1L, 4L, 2L, 1L, 2L, 3L, 1L, 4L), .Label = c("Default", 
"Visual element", "Verbal content", "Visual design", "Combined", 
"DesignZH"), class = "factor"), ISFregscores = c(0.984372106429775, 
-0.383676865152824, -0.816194838031774, -0.408554787302724, -0.0416530380928891, 
0.998088756156888, 0.216609251327447, 0.83416518546863, 1.00178246600492, 
-0.496215251116934, -1.34559758838579, NA, 0.707838661016661, 
1.05815783619489, -0.314855036376305, 0.617674358967702, -0.56862344822269, 
0.0589354712707628, 0.31998903974822, -0.511084756816837, -0.171121724458495, 
0.532699047600051, 0.196311893993997, -2.09902298349596, 1.04422334581248, 
-0.132687312769232, 1.05733961165571, 0.541606480874359, 0.440296538856025, 
0.895064902672922), age_dup = structure(c(2L, 1L, 1L, 2L, 1L, 
2L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 3L, 1L, 2L, 3L, 
1L, 2L, 1L, 2L, 2L, 2L, 1L, 2L, 1L), .Label = c("under34", "age35_49", 
"over50"), class = "factor"), edu = structure(c(5L, 4L, 5L, 2L, 
5L, 5L, 5L, 5L, 5L, 4L, 5L, NA, 5L, 4L, 1L, 5L, 4L, 5L, 3L, 5L, 
2L, 5L, 3L, 5L, 6L, 5L, 6L, 1L, 3L, 5L), .Label = c("oblig. Schulzeit", 
"Berufsausbildung", "Berufsmatura", "Gymnasiale Matura", "BA/MA", 
"Doktorat", "Andere"), class = "factor"), empl = structure(c(1L, 
6L, 1L, 2L, 8L, 2L, 5L, 2L, 2L, 6L, 2L, NA, 1L, 1L, 6L, 2L, 6L, 
1L, 1L, 4L, 2L, 6L, 1L, 1L, 3L, 6L, 2L, 2L, 4L, 1L), .Label = c("Privatsektor", 
"öffentlicher Sektor", "Non-Profit Sektor", "selbstständig", 
"Rentner/in", "Student/in", "Hausfrau/Hausmann", "arbeitssuchend"
), class = "factor"), civ_dup = structure(c(2L, 1L, 1L, 3L, 2L, 
1L, 2L, 2L, 2L, 1L, 2L, NA, 2L, 2L, 1L, 1L, 1L, 2L, 1L, 3L, 1L, 
2L, 2L, 2L, 2L, 1L, 2L, 3L, 2L, 1L), .Label = c("single", "Partnerschaft", 
"keine Angabe"), class = "factor"), kids = structure(c(2L, 1L, 
1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, NA, 2L, 2L, 1L, 1L, 1L, 1L, 
2L, 2L, 1L, 1L, 1L, 2L, 2L, 1L, 1L, 1L, 1L, 1L), .Label = c("Nein", 
"Ja"), class = "factor"), known = structure(c(1L, 1L, 1L, 1L, 
1L, 1L, 1L, 2L, 2L, 1L, 1L, NA, 1L, 2L, 1L, 1L, 1L, 2L, 1L, 1L, 
1L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 2L, 1L), .Label = c("Nein", "Ja"
), class = "factor"), device = structure(c(1L, 2L, 1L, 2L, 2L, 
1L, 3L, 1L, 2L, 2L, 1L, NA, 2L, 1L, 1L, 1L, 2L, 2L, 1L, 1L, 2L, 
1L, 1L, 1L, 1L, 2L, 2L, 1L, 1L, 2L), .Label = c("Smartphone / Tablet iOS (iPhone/iPad)", 
"Smartphone / Tablet (Android)", "Computer / Laptop"), class = "factor")), row.names = c(NA, 
-30L), class = c("tbl_df", "tbl", "data.frame"))

And this is the code for my regression that I then use for the interaction (graph):

Model <- lm(dv ~ gend * group + ISFregscores + age_dup + edu + empl + civ_dup + kids + known + device, data=)

Can you make your post [reproducible](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) by providing your data using `dput()`? — jrcalabrese, Dec 12 '22 at 16:44

dipetkov · Accepted Answer · 2023-01-11T15:46:21.810

You've got the right approach to change the font but you also have to make sure the font is actually available to the graphics device. This step can be tricky; I use the showtext package which makes this a bit easier.

To change the color palette, specify the color scale (rather than the fill scale).

library("showtext")
#> Loading required package: sysfonts
#> Loading required package: showtextdb
library("emmeans")
library("tidyverse")
showtext_auto()

# 30 data points are too few to fit the original model, so I drop `device`
model <- lm(
  dv ~ gend * group + ISFregscores + age_dup + edu + empl + civ_dup + kids + known,
  data = data
)

p <- emmip(
  model, group ~ gend,
  CIs = TRUE,
  nuisance = c("known", "age_dup", "edu"),
  xlab = "",
  ylab = "Intention to use the platform"
)
p +
  scale_color_brewer(
    palette = "Blues"
  ) +
  guides(
    color = guide_legend(title = "New Legend Title")
  ) +
  theme(
    text = element_text(family = "serif", face = "bold.italic", size = 16)
  )

^{Created on 2023-01-11 with reprex v2.0.2}

Do you also maybe have a hint on how I could modify the legend from "group" into "Group". I already tried the following but it didn't work: `p + labs(fill = "Group")` or `p + scale_fill_discrete(name = "Group", labels = c("Default", "Visual Element", "Verbal content", "Visual design", "Combined")) ` — phdmeme, Jan 11 '23 at 15:38
Sure. One way is to add: `+ guides(color = guide_legend(title="New Legend Title"))`. Let me update the example so you see it in action. — dipetkov, Jan 11 '23 at 15:45

Modify color and font of emmip plot (emmeans package) in R

1 Answers1