Search code examples
rggplot2bar-chartoffsetcolumn-chart

Add vertical offset to stacked column plot with ggplot2


I have several stacked column charts representing drilling profiles. I want to offset the y-position of each Borehole to represent the actual height on the ground.

My Data looks like this:

 x layer.thickness layer.depth Petrography    BSCategory  Offset
 0             0.2         0.2        silt     Drilling1      0
 0             1.0         1.2      gravel     Drilling1      0
 0             3.0         4.2        silt     Drilling1      0
 4             0.4         0.4        silt     Drilling2     -1
 4             0.8         1.2      gravel     Drilling2     -1
 4             2.0         3.2        sand     Drilling2     -1

My minimum working code so far is this:

df <- data.frame(x=c(0,0,0,4,4,4), layer.thickness = c(0.2,1.0,3.0,0.4,0.8,2.0), 
                 layer.depth = c(0.2,1.2,4.2,0.4,1.2,3.2), 
                 Petrography = c("silt", "gravel", "silt", "silt", "gravel", "sand"),
                 BSCategory = c("Drilling1","Drilling1","Drilling1","Drilling2","Drilling2","Drilling2"),
                 Offset = c(0,0,0,-1,-1,-1))

# provide a numeric ID that stops grouping individual petrography items
df <- transform(df,ix=as.numeric(factor(df$BSCategory)));


drilling <- ggplot(data = df, aes(x = x, y = layer.thickness, group = ix, fill = Petrography)) +
  theme_minimal() + 
  theme(axis.line = element_line(colour = "black"),
        panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.border = element_blank(),
        panel.background = element_blank(),
        axis.line.x = element_blank(),
        axis.ticks.y = element_line(),
        aspect.ratio=1) +
  geom_col(position = position_stack(reverse = TRUE), width= .15,color="black") +
  scale_y_reverse(expand = c(0, 0), name ="Depth [m]") +
  scale_x_continuous(position = "top", breaks = df$x, labels=paste(df$BSCategory,"\n",df$x,"m"), name="") +
  scale_fill_manual(values = c("gravel"='#f3e03a', "sand"= '#e09637', "silt"='#aba77d'))

print(drilling)

This is my output so far (with red indicating what it should look like):

Example of drilling profiles


Solution

  • It may be easier to work with geom_rect here (bar charts are/should be anchored at zero). First we need to calculate y start and end positions for each sample:

    library(data.table)
    setDT(df)[ , `:=`(start = start <- c(Offset[1] + c(0, cumsum(head(layer.thickness, -1)))),
                      end = start + layer.thickness), by = BSCategory]
    

    geom_rect has both fill and color aesthetics, which makes it easy to add a border to each individual sample.

    w <- 0.5 # adjust to desired width
    
    ggplot(data = df, aes(xmin = x - w / 2, xmax = x + w / 2, ymin = start, ymax = end,
                          fill = Petrography, group = Petrography)) +
      geom_rect(color = "black") +
      scale_y_reverse(expand = c(0, 0), name ="Depth [m]") +
      scale_x_continuous(position = "top", breaks = df$x, labels = paste(df$BSCategory,"\n", df$x, "m"), name = "") +
      scale_fill_manual(values = c("gravel" = '#f3e03a', "sand" = '#e09637', "silt" = '#aba77d')) +
      theme_classic() +
      theme(axis.line.x = element_blank(),
            axis.ticks.x = element_blank())
    

    enter image description here


    Alternatively, geom_segment could be used. geom_segment doesn't have a fill aes, so we need to change to color:

    ggplot(data = df, aes(x = x, xend = x, y = start, yend = end, color = Petrography, group = ix)) +
      geom_segment(size = 5) +
      scale_y_reverse(expand = c(0, 0), name ="Depth [m]") +
      scale_x_continuous(position = "top", breaks = df$x, labels = paste(df$BSCategory,"\n", df$x, "m"), name = "") +
      scale_color_manual(values = c("gravel" = '#f3e03a', "sand" = '#e09637', "silt" = '#aba77d')) +
      theme_classic() +
      theme(axis.line.x = element_blank(),
            axis.ticks.x = element_blank())
    

    enter image description here

    To add borders, see Add border to segments in geom_segment.