rggplot2

How to adjust the width of my boxplot without affecting my scatterplot points

I need some help with a figure. I need a boxplot in which the y axis is total number of brain cells, and x is ordered by mean body mass, grouped by species and is coloured according to the sociality.

Please find the R code I am using and the data below:

ggplot(Data, aes(x=mean_bodymass, y = Total.brain.cell, group=Species))+
  geom_boxplot (aes(fill=Sociality),width=5)+
  geom_point (aes(fill=Sociality),shape=21,size=4)+
scale_y_continuous(trans='log10', n.breaks = 8,labels = comma)+
  scale_x_continuous(trans='log10',labels = comma)+
theme_classic()+
scale_fill_manual(values=c("#FF1493","#BF3EFF","#FF3030", "#FF7F00","#0066ff"))+
  guides(fill="none")+
  theme(base_size=10, text = element_text(size = 30))+
  labs(x = "body mass(g)(log)", y = "total brain cells(log)")

Small sample of my data:

Data <- structure(list(Sociality = c("Semi-social", "Semi-social", "Semi-social", 
"Solitary", "Solitary", "Solitary", "Solitary", "Solitary", "Solitary", 
"Solitary", "Solitary", "Solitary", "Solitary", "Solitary", "Solitary", 
"Solitary", "Solitary", "Solitary", "Kleptoparasitism", "Kleptoparasitism", 
"Kleptoparasitism"), Species = c("Lipotriches australica", "Lipotriches australica", 
"Lipotriches australica", "Hylaeus  euxanthus", "Hylaeus  euxanthus", 
"Hylaeus  euxanthus", "Leioproctus amabilis", "Leioproctus amabilis", 
"Leioproctus amabilis", "Leioproctus recusus", "Leioproctus recusus", 
"Leioproctus recusus", "Megachile (Hacteriapis)", "Megachile (Hacteriapis)", 
"Megachile (Hacteriapis)", "Pseudoanthidium repetitum", "Pseudoanthidium repetitum", 
"Pseudoanthidium repetitum", "Thyreus nitidulus", "Thyreus nitidulus", 
"Thyreus nitidulus"), Total.brain.cell = c(1468750, 1441250, 
1496250, 906250, 853750, 870000, 961250, 975000, 947500, 855000, 
958750, 988750, 1952500, 1926250, 1993750, 1411250, 1075000, 
1450000, 1781250, 1757500, 1742500), mean_bodymass = c(0.065933333, 
0.065933333, 0.065933333, 0.008333333, 0.008333333, 0.008333333, 
0.0377, 0.0377, 0.0377, 0.027433333, 0.027433333, 0.027433333, 
0.053966667, 0.053966667, 0.053966667, 0.043, 0.043, 0.043, 0.084533333, 
0.084533333, 0.084533333)), row.names = 4:24, class = "data.frame")

My issue is the width of the boxplot. This is the result without adjusting the width of the boxplot:

enter image description here

And this is the result when adjusting the boxplot width (see code below), where the points are not not inline with the boxplot.

enter image description here

Is there any way I can adjust the width of the boxplot without affecting the points? Or do I need to do two separate plots and combine them? Any help would be appreciated.

Solution

  • Note that the default action is geom_boxplot(position="dodge2"), which means it uses position_dodge2(). I couldn't find a width= greater than width=0.05 that kept the points aligned with the boxplots, it always nudged them out of the way.

    One alternative is to use position_dodge() instead.

    Original,

      geom_boxplot(aes(fill=Sociality), width=5) +

    original boxplot using position_dodge2()

    Shifting to position_dodge() and reducing the width:

      geom_boxplot(aes(fill=Sociality), position=position_dodge(), width=0.1) +

    same boxplot using position_dodge() and width of 0.1

    position_dodge() allows for overlaps, you can get fairly crazy with it. Because your boxplots are positioned on a continuous scale and you have unique values that are relatively close to each other, you may have to play with the width= to get them exactly how you like.