Reference Color from Data Frame in ggplot2

Question

MWE:

library(ggplot2)
library(dplyr)

df <- data.frame(GROUP = factor(c("AL", "AS", "B", "HI", "I", "P", "T", "W", "E", "NE", "EL", "EP", "SW", "SOD"), ordered = TRUE),
                 color.codes = c("#000000", 
                                 "#b05cc6",
                                 "#afb045",
                                 "#7777c8",
                                 "#db8f47",
                                 "#4dacd0",
                                 "#cb5242",
                                 "#56b88f",
                                 "#c75a88",
                                 "#5db74e",
                                 "#4f7e3b",
                                 "#ff0000",
                                 "#8d7134",
                                 "#000080"),
                 Year = rep(paste0(2010:2025, "-", 2011:2026), each=14),
                 value = runif(224))

df$Year <- factor(df$Year,
                  levels=sort(unique(df$Year)), 
                  ordered=TRUE)

ggplot(df, 
               aes(x = Year, y = value, color = GROUP, group = GROUP)) + 
  geom_point() + 
  geom_line(linetype = "dotted") + # make all dotted by default
  geom_line(data = (df %>% filter(Year <= "2014-2015")),
            aes(x = Year, y = value, color = GROUP),
            lty = "solid") + # and then make all prior to 2015 solid
  scale_color_manual(values = df$color.codes) +
  xlab("School Year") + 
  ylab("Proficiency Rate") 

Let's examine the data frame:

> head(df)
  GROUP color.codes      Year     value
1    AL     #000000 2010-2011 0.5540173
2    AS     #b05cc6 2010-2011 0.0301139
3     B     #afb045 2010-2011 0.7469069
4    HI     #7777c8 2010-2011 0.4981707
5     I     #db8f47 2010-2011 0.9677573
6     P     #4dacd0 2010-2011 0.7777855

For example, #b05cc6, according to Google, is something of a purple shade. But it clearly looks red, despite that it's supposed to be purple.

It also doesn't make any sense that B has the same color as AL. How can this be fixed?

Answer 1

Make a palette of group names and colour code mappings:

pal = unique(df[,c("GROUP","color.codes")])

> head(pal)
  GROUP color.codes
1    AL     #000000
2    AS     #b05cc6
3     B     #afb045
4    HI     #7777c8
5     I     #db8f47
6     P     #4dacd0

Fixup color codes to be character and not factor, or else they get converted to numeric and looked up in the palette:

> pal$color.codes=as.character(pal$color.codes)

Now create a named vector from the palette columns:

> pval = pal$color.codes
> names(pval) = pal$GROUP
> pval
       AL        AS         B        HI         I         P         T         W 
"#000000" "#b05cc6" "#afb045" "#7777c8" "#db8f47" "#4dacd0" "#cb5242" "#56b88f" 
        E        NE        EL        EP        SW       SOD 
"#c75a88" "#5db74e" "#4f7e3b" "#ff0000" "#8d7134" "#000080" 

Then use that in scale_color_manual(values=pval).

Now AS looks purply, B looks like a sort of greeny vom colour.

EP looks red. etc.