This page explains how to build a dotplot histogram with R, ggplot2 and plotly. This type of visualization in a specific type of histogram. It shows the distribution of a numeric variable. But instead of using bars, each individual observation is represented as a dot.

It particularly makes sense to use interactivity for dotplot histograms: hovering a datapoint will give you more information about its identity.

The idea is to split our numerical variable in several bins, and to calculate the position on the Y axis for each individual unit. Once this new information is available, it is possible to use the geom_point like if it was a scatterplot.


# Library
library(tidyverse)
library(plotly)

# A classic histogram for the iris data set (left)
ggplot(iris, aes(x=Sepal.Length)) +
  geom_histogram()

# Transform a litte bit the dataset to make dots
don = iris %>% 
  arrange(Sepal.Length) %>% # sort using the numeric variable that interest you
  mutate(var_rounded = (Sepal.Length+1) - ( (Sepal.Length+1) %% 0.2 ) ) %>% # This attributes a bin to each observation. Here 0.2 is the size of the bin.
  mutate(y=ave(var_rounded, var_rounded, FUN=seq_along)) # This calculates the position on the Y axis: 1, 2, 3, 4...

# Make the plot (middle)
ggplot(don, aes(x=var_rounded, y=y) ) +
  geom_point( size=6, color="skyblue" ) 

# Improve the plot, and make it interactive (right)
don=don %>% mutate(text=paste("ID: ", rownames(iris), "\n", "Sepal Length: ", Sepal.Length, "\n", "Species:: ", Species, sep="" )) 
p=ggplot(don, aes(x=var_rounded, y=y) ) +
  geom_point( aes(text=text), size=6, color="skyblue" ) +
  xlab('Sepal Length') +
  ylab('# of individual') +
  theme_classic() +
  theme(
    legend.position="none",
    axis.line.y = element_blank(),
    axis.text=element_text(size=15)
  )
p

# Use the magic of ggplotly to have an interactive version
ggplotly(p, tooltip="text")

# Library

library(tidyverse)

library(plotly)

# A classic histogram for the iris data set (left)

ggplot(iris, aes(x=Sepal.Length)) +

geom_histogram()

# Transform a litte bit the dataset to make dots

don = iris %>%

arrange(Sepal.Length) %>% # sort using the numeric variable that interest you

mutate(var_rounded = (Sepal.Length+1) - ( (Sepal.Length+1) %% 0.2 ) ) %>% # This attributes a bin to each observation. Here 0.2 is the size of the bin.