Two-factor bar plot

Question

Working from the following dataframe:

    > foo
        species density day percent  
    1  species1    high   1    0.40  
    2  species1     low   1    0.20  
    3  species1  medium   1    0.40  
    4  species2    high   1    0.35  
    5  species2     low   1    0.10  
    6  species2  medium   1    0.55  
    7  species3    high   1    0.35  
    8  species3     low   1    0.20  
    9  species3  medium   1    0.45  
    10 species4    high   1    0.30  
    11 species4     low   1    0.20  
    12 species4  medium   1    0.50  
    13 species1    high 100    0.50  
    14 species1     low 100    0.40  
    15 species1  medium 100    0.10  
    16 species2    high 100    0.40  
    17 species2     low 100    0.05  
    18 species2  medium 100    0.55  
    19 species3    high 100    0.65  
    20 species3     low 100    0.05  
    21 species3  medium 100    0.30  
    22 species4    high 100    0.40  
    23 species4     low 100    0.20  
    24 species4  medium 100    0.40  

I have created the following faceted bar graph:

require(ggplot2)

foo$density<-factor(foo$density,levels=c('low','medium','high'))

d <- ggplot(foo, aes(x=species, y=percent, fill=density)) +
    geom_bar(aes(width=.65), stat="identity") +
    facet_grid(. ~ day)

However, I would like to merge these graphs to create a single two-factor bar graph. On the x-axis, each day- 1 and 100 - would be grouped by species. Any suggestions on how to create this?

Many thanks!

Answer 1

Here is another approach that incorporates @Joran's suggestion along with your requirements.

I have changed day to a factor, but also changed the default x-axis label to "Species". Also, because density is clearly a sequential factor, I have use a sequential color scale from Color Brewer (http://colorbrewer2.org/). You can experiment with color scales by changing palette number or calling palette by name scale_fill_brewer(palette="GnBu"). Find the palette names on the Color Brewer web page.

foo <- read.table(header=TRUE,
                 text="species  density day percent  
                    1  species1    high   1    0.40  
                    2  species1     low   1    0.20  
                    3  species1  medium   1    0.40  
                    4  species2    high   1    0.35  
                    5  species2     low   1    0.10  
                    6  species2  medium   1    0.55  
                    7  species3    high   1    0.35  
                    8  species3     low   1    0.20  
                    9  species3  medium   1    0.45  
                    10 species4    high   1    0.30  
                    11 species4     low   1    0.20  
                    12 species4  medium   1    0.50  
                    13 species1    high 100    0.50  
                    14 species1     low 100    0.40  
                    15 species1  medium 100    0.10  
                    16 species2    high 100    0.40  
                    17 species2     low 100    0.05  
                    18 species2  medium 100    0.55  
                    19 species3    high 100    0.65  
                    20 species3     low 100    0.05  
                    21 species3  medium 100    0.30  
                    22 species4    high 100    0.40  
                    23 species4     low 100    0.20  
                    24 species4  medium 100    0.40")

foo$density <- factor(foo$density, levels=c("low", "medium", "high"))
foo$day <- factor(paste("Day", foo$day, sep="_"))

library(ggplot2)

d2 <- ggplot(foo, aes(x=day, y=percent, fill=density)) +
      theme_bw() +
      geom_bar(width=0.95, stat="identity") +
      scale_fill_brewer(type="seq", palette=15) +
      xlab("Species") +
      opts(axis.text.x=theme_text(size=6)) +
      facet_grid(. ~ species)

ggsave("barplot_1.png", d2, width=6, height=4)

One problem I have not solved is that the levels of density do not stack in the proper order (for me or in @MYaseen208's answer). Stacking is correct in the original post. Does anyone know what the problem is?

Answer 2

Try this one

foo$species_day <- with(data = foo, expr = paste(species, day))
d <-ggplot(foo, aes(x=species_day, y=percent, fill=density)) +
      geom_bar(aes(width=.65), stat="identity")

You can rearrange the levels if you want.