<dot plots with continuous 3 variables and 4 variables>
As you can easily understand, the variables x and y on x and y-axis are well represented even if they are continuous.
Then what about adding another continuous variable to the dot plot?
You can make a dot plot with smoothly colored dots that express the new continuous variable.
Also, You can make a dot plot with smoothly getting bigger dots that express another continuous variable.
For this post, I will use the 'USJudgeRatings' data in R.
> ?USJudgeRatings
<Set a continuous variable and distinguish dots by filling continuous colors>
For this part, I will use the variable 'PREP' which consists of continuous elements for filling the dots.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,colour=PREP))+geom_point()
You can adjust the shape of dots in the plot by using geom_point(shape, size)
Also, you can adjust the colour to fill the continuous variable by setting scale_fill_gradient(low="",high="")
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP,group=PREP)) +geom_point(shape=21,size=3) +scale_fill_gradient(low="black",high="white") |
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP) +geom_point(shape=21,size=3) +scale_fill_gradient(low="white",high="black") |
|
|
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP,group=PREP)) +geom_point(shape=21,size=3) +scale_fill_gradient(low="white",high="black") |
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP,group=PREP)) +geom_point(shape=21,size=3) +scale_fill_gradient(low="pink",high="black") |
|
|
You can concretely set the legend beside the plot.
legend - <5,6,7,8,9>
The variable 'PREP' consists of elements between 5 to 9.5. So the basic legend has breaks on integers 5,6,7,8,9.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP,group=PREP))+geom_point(shape=21,size=3)+scale_fill_gradient(low="white",high="red")
legend - <5, 6.5, 8>
The variable 'PREP' consists of elements between 5 to 9.5 and I set the breaks of the legend as 2.5 different for each other.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP))+geom_point(shape=21,size=3)+scale_fill_gradient(low="white",high="red",breaks=seq(5,10,by=1.5))
legend - <to be expressed as discrete>
The variable 'PREP' consists of elements between 5 to 9.5. So the basic legend has breaks on integers 5,6,7,8,9. The only difference with the first example is that this expresses the colors by separating the continous color bar.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP))+geom_point(shape=21,size=3)+scale_fill_gradient(low="white",high="red",guide=guide_legend())
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,fill=PREP))+geom_point(shape=21,size=3)+scale_fill_gradient(low="white",high="red",breaks=seq(5,10,by=1.5),guide=guide_legend())
<set a continuous variable and distinguish dots by setting different sizes>
dots in the dot plot is originally set as having sizes between 1 and 6.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=PREP))+geom_point()
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=PREP))+geom_point()+scale_size_continuous(range=c(0,4))
scale_size_area() is needed to set the size of dots proportional to numeric elements.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=PREP))+geom_point()+scale_size_area()
Setting alpha is the usual task when making a dot plot because of overlapped dots.
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=PREP))+geom_point(alpha=0.6)+scale_size_area()
<set two continuous variables and distinguish dots by setting different sizes and colors>
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=DECI,fill=PREP))+geom_point(shape=21)+scale_size_area()+scale_fill_gradient(low="white",high="blueviolet")
> ggplot(data=USJudgeRatings,aes(x=DILG,y=FAMI,size=DECI,fill=PREP))+geom_point(shape=21)+scale_size_continuous(range=c(0,6))+scale_fill_gradient(low="white",high="blueviolet")
