Assign colours in ggplot from a manual value

Question

I have a df, df_Filtered, that looks like this:

Country Region  Sales Year Colour   
Germany Berlin  2000  2000 #FF0000
Germany Hamburg 1500  2001 #33CC33
Germany Kiel    2150  2002 #00FF00
UK      London  1200  2000 #CC0000
UK      York    1300  2001 #FFFF33
UK      Leeds   2000  2002 #339900
Japan   Tokyo   500   2000 #66CC00
Japan   Kyoto   750   2001 #990099

I want to plot the data for each region and year wrt the sales value:

ggplot(df_Filtered, aes(x = Year, y = Sales, colour = Region, scale_y_continuous(breaks = 1), size=mysize, labels=as.matrix(df_Filtered_regions))) +  geom_line() +
    labs(x = "Years", y = "Sales", title = NULL) + 
    scale_x_continuous(breaks = c(2000, 2001, 2002)) + 
    scale_size(range = c(1, 4), guide="none") +
    theme(panel.background = element_blank()) + 
    theme(legend.position="bottom") +
    scale_color_discrete(name=NULL) +
    theme(plot.background = element_rect(colour = "black", size = 1)) +
    theme(axis.line = element_line())

Note that I also have a vector for the size of the lines that I want to use in the real data. In the real code the lines representing the average country value is thicker, but to make it simple:

Size = 2
df_Filtered$mysize <- rep(Size, nrow(df_Filtered))

I want to use the colours from the column Colour in df_Filtered for the lines in the plot. How do I do this?

(I previously tried to assign the colors a by an automatic process, but that did not work out very well: Different colour palettes for different series ggplot )

Answer 1

Please make an effort in future questions to provide working code, esp since the ggplot2 construct is seriously incorrect in many, many places. I also strongly suggest you do not use the colors you've chosen for the final plot. Finally, decimal ascii code character 32 (i.e " ") is free and I suspect adopting and adhering to a code format style would help avoid some of the plot construct errors in the future.

You can use pre-generated geom-component colors using I() which marks the column "asis" which informs ggplot2 to just use the values specified.

library(ggplot2)

read.csv(text="Country,Region,Sales,Year,Colour
Germany,Berlin,2000,2000,#FF0000
Germany,Hamburg,1500,2001,#33CC33
Germany,Kiel,2150,2002,#00FF00
UK,London,1200,2000,#CC0000
UK,York,1300,2001,#FFFF33
UK,Leeds,2000,2002,#339900
Japan,Tokyo,500,2000,#66CC00
Japan,Kyoto,750,2001,#990099", stringsAsFactors = FALSE) -> xdf

xdf$mysize <- rep(2, nrow(xdf))

ggplot(xdf) +  
  geom_line(
    aes(
      x = Year, y = Sales, group = Country,
      colour = I(Colour), size = mysize
    )
  ) +
  scale_x_continuous(breaks = c(2000, 2001, 2002)) + 
  scale_y_continuous(breaks = 1) +
  scale_size(range = c(1, 4), guide = "none") +
  labs(x = "Years", y = "Sales", title = NULL) + 
  theme(axis.line = element_line()) +
  theme(panel.background = element_blank()) + 
  theme(plot.background = element_rect(colour = "black", size = 1)) +
  theme(legend.position = "bottom")

ggplot2 plotting idioms get easier to grok as you use them. Practice is essential.

You want a legend, but you didn't work through how to use a manual color scale. We'll build a named vector from the data you specified:

manual_scale_colors <- setNames(xdf$Colour, xdf$Region)

then remove the I() and add back the manual scale call:

ggplot(xdf) +  
  geom_line(
    aes(
      x = Year, y = Sales, group = Country,
      colour = Region, size = mysize
    )
  ) +
  scale_x_continuous(breaks = c(2000, 2001, 2002)) + 
  scale_y_continuous(breaks = 1) +
  scale_color_manual(values = manual_scale_colors) +
  scale_size(range = c(1, 4), guide = "none") +
  labs(x = "Years", y = "Sales", title = NULL) + 
  theme(axis.line = element_line()) +
  theme(panel.background = element_blank()) + 
  theme(plot.background = element_rect(colour = "black", size = 1)) +
  theme(legend.position = "bottom")