Statistics Sunday: Creating a Stacked Bar Chart for Rank Data

At work on Friday, I was trying to figure out the best way to display some rank data. What I had were rankings from 1-5 for 10 factors considered most important in a job (such as Salary, Insurance Benefits, and the Opportunity to Learn), meaning each respondent chose and ranked the top 5 from those 10, and the remaining 5 were unranked by that respondent. Without even thinking about the missing data issue, I computed a mean rank and called it a day. (Yes, I know that ranks are ordinal and means are for continuous data, but my goal was simply to differentiate importance of the factors and a mean seemed the best way to do it.) Of course, then we noticed one of the factors had a pretty high average rank, even though few people ranked it in the top 5. Oops.

So how could I present these results? One idea I had was a stacked bar chart, and it took a bit of data wrangling to do it. That is, the rankings were all in separate variables, but I want them all on the same chart. Basically, I needed to create a dataset with:

1 variable to represent the factor being ranked
• 1 variable to represent the ranking given (1-5, or 6 that I called “Not Ranked”)
• 1 variable to represent the number of people giving that particular rank that particular factor

What I ultimately did was run frequencies for the factor variables, turn those frequency tables into data frames, and merged them together with rbind. I then created chart with ggplot. Here’s some code for a simplified example, which only uses 6 factors and asks people to rank the top 3.

First, let’s read in our sample dataset – note that these data were generated only for this example and are not real data:

library(tidyverse)

## -- Attaching packages --------------------------------------------------------------------------------------------------------------------- tidyverse 1.2.1 --

## v ggplot2 3.0.0     v purrr   0.2.4
## v tibble  1.4.2     v dplyr   0.7.4
## v tidyr   0.8.0     v stringr 1.3.1
## v readr   1.1.1     v forcats 0.3.0

## Warning: package 'ggplot2' was built under R version 3.5.1

## -- Conflicts ------------------------------------------------------------------------------------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

ranks <- read_csv("C:/Users/slocatelli/Desktop/sample_ranks.csv", col_names = TRUE)

## Parsed with column specification:
## cols(
##   RespID = col_integer(),
##   Salary = col_integer(),
##   Recognition = col_integer(),
##   PTO = col_integer(),
##   Insurance = col_integer(),
##   FlexibleHours = col_integer(),
##   OptoLearn = col_integer()
## )

This dataset contains 7 variables – 1 respondent ID and 6 variables with ranks on factors considered important in a job: salary, recognition from employer, paid time off, insurance benefits, flexible scheduling, and opportunity to learn. I want to run frequencies for these variables, and turn those frequency tables into a data frame I can use in ggplot2. I’m sure there are much cleaner ways to do this (and please share in the comments!), but here’s one not so pretty way:

salary <- as.data.frame(table(ranks$Salary))
salary$Name <- "Salary"
recognition <- as.data.frame(table(ranks$Recognition))
recognition$Name <- "Recognition by \nEmployer"
PTO <- as.data.frame(table(ranks$PTO))
PTO$Name <- "Paid Time Off"
insurance <- as.data.frame(table(ranks$Insurance))
insurance$Name <- "Insurance"
flexible <- as.data.frame(table(ranks$FlexibleHours))
flexible$Name <- "Flexible Schedule"
learn <- as.data.frame(table(ranks$OptoLearn))
learn$Name <- "Opportunity to \nLearn"

rank_chart <- rbind(salary, recognition, PTO, insurance, flexible, learn)
rank_chart$Var1 <- as.numeric(rank_chart$Var1)

With my not-so-pretty data wrangling, the chart itself is actually pretty easy:

ggplot(rank_chart, aes(fill = Var1, y = Freq, x = Name)) +
  geom_bar(stat = "identity") +
  labs(title = "Ranking of Factors Most Important in a Job") +
  ylab("Frequency") +
  xlab("Job Factors") +
  scale_fill_continuous(name = "Ranking",
                      breaks = c(1:4),
                      labels = c("1","2","3","Not Ranked")) +
  theme_bw() +
  theme(plot.title=element_text(hjust=0.5))