0

The data frame has many continuous numeric columns (e.g. gr) and a sample identifier - wellseq. There are many rows of data for each wellseq. In a data frame - there are 94 distinct levels of wellseq in 10227 rows. Header lines from data frame are:

 gr wellseq
 1 27.7049       1
 2 31.1149       1
 3 34.5249       1
 4 39.7249       1
 5 44.9249       1
 6 50.1299       1

Summary of column gr is as below:

summary(GR)
       gr            
 Min.   :-6.94     
 1st Qu.:10.71   
 Median :13.76   
 Mean   :18.99   
 3rd Qu.:20.70   
 Max.   :98.42   
 NA's   :55

Basic Histogram of the entire data for the gr is suitably created. For further analysis, it is required to identify each wellseq contributing in the histogram. The ggplot() script used is:

p2 <- ggplot() + theme_bw() + 
        geom_histogram(data=GR, na.rm= TRUE, mapping = aes(x=gr, fill=factor(GR$wellseq)), 
           bins = 10) + scale_color_brewer(palette = "Dark2") +
        scale_x_continuous(limits = c(-10, 100)) +
        labs(title=paste("Gamma Ray","Histogram", sep=" ")) +
        theme(legend.position = "none")

Sequential Color in Histogram The resulting output has color - which is "sequential" and NOT the "qualitative" palette "Dark2". I tried using the answer in "How to generate a number of most distinctive colors in R?" @ stackoverflow.com and created required colors.

Dcolor = grDevices::colors()[grep('gr(a|e)y', grDevices::colors(), invert = T)]
DcolorR <- sample(Dcolor, 433, replace = F)

using scale_colour_manual(values = DcolorR)

gives the same histogram. Using ..count.. for y the histogram does show the boundaries for different wellseq but does not fill as needed. 

p3 <- ggplot() + theme_bw() +
    geom_histogram(data=GR, na.rm= TRUE, mapping = aes(x=gr, y= ..count.., col = factor(GR$wellseq), bins = 10)) +
    scale_colour_manual(values = DcolorR) +
    scale_x_continuous(limits = c(-10, 100)) +
    labs(title=paste("Gamma Ray"," Frequency Histogram", sep=" ")) +
    theme(legend.position = "none")
fill = 1 # leads to blue colored staked histogram

..count.. creates discrete boundaries in the bars. Fill is single color I am trying to get a plot like attached. Please guide. Thanks in advance.

Desired Output

Sumedh
  • 4,345
  • 1
  • 15
  • 31
Srikanth G
  • 91
  • 2
  • 8
  • 3
    you need to use `fill=` not `col=` to control the fill color. Then change `scale_color_manual` to `scale_fill_manual`.Also, please see [how to make a great R reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). – C8H10N4O2 Aug 01 '16 at 16:33

1 Answers1

0

If you set aes(x=gr, fill=wellseq) then you should get what you are looking for which is a grouping of the subsets of gr as defined by their membership in wellseq.

Look here.Previous simple version of a grouped histogram R histogram with multiple populations

Community
  • 1
  • 1
sconfluentus
  • 3,737
  • 1
  • 18
  • 38