55  Frequency Plot (Bar Plot)

55.1 Definition

The Frequency Plot displays the absolute frequencies or counts for different items or categories. This plot is used with qualitative data - in case of quantitative data one should use the histogram instead (cf. Chapter 62). Even though it is not required, we display the absolute frequencies in sorted order -- this allows the user to easily see the ranking of each category.

55.2 Horizontal axis

The horizontal axis displays the categories of the variable under investigation. If there is sufficient space, the labels of each category are printed along the axis. The categories are ordered according to their absolute frequencies (from high to low). The width of the bars that are displayed have no meaning.

55.3 Vertical axis

The counts (i.e. absolute frequencies) of each category are represented by the vertical axis. The height of the bars that are displayed correspond to the absolute frequencies on the vertical axis.

55.4 R Module

The Frequency Plot can be found on the public website:

Note: Frequency Plots and Histograms are generated by the same R Module (even though both methods do not have the same interpretation). The R Module automatically detects whether the data are qualitative or quantitative (the appropriate method is selected by the software).

The Frequency plot is available in RFC under the menu “Descriptive / Histogram & Frequency Table”.

If you prefer to compute the Frequency Plot on your local computer, the following code snippet can be used in the R console:

Code 
x <- c('Firefox','MSIE','MSIE','MSIE','MSIE','Firefox','Firefox','MSIE','Chrome','Firefox','Firefox','MSIE','MSIE','MSIE','MSIE','MSIE','Firefox','MSIE','Firefox','Firefox','MSIE','Firefox','Firefox','MSIE','MSIE','Firefox','Firefox','Safari','MSIE','Firefox','MSIE','MSIE','MSIE','Firefox','MSIE','MSIE','Firefox','MSIE','MSIE','MSIE','Firefox','MSIE','MSIE','Firefox','Opera','Firefox','MSIE','MSIE','MSIE','Firefox','Firefox','Safari','Firefox','Firefox','MSIE','MSIE','MSIE','Firefox','Safari','MSIE','MSIE','MSIE','MSIE','Firefox','MSIE','Chrome','MSIE','Firefox','Safari','Safari','Firefox','MSIE','MSIE','Firefox','MSIE','MSIE','MSIE','MSIE','Firefox','MSIE','MSIE','MSIE','MSIE','Firefox','MSIE','Firefox','Safari','Firefox','MSIE','MSIE','MSIE','MSIE','MSIE','MSIE','MSIE','Safari','Firefox','MSIE','Firefox','MSIE','MSIE','MSIE','MSIE','MSIE','Firefox','MSIE','Firefox','MSIE','MSIE','MSIE','Firefox','MSIE','MSIE','MSIE','Safari','MSIE','Firefox','Firefox','MSIE','Safari','Safari','Firefox','MSIE','MSIE','MSIE','MSIE','Chrome','MSIE','Firefox','MSIE','MSIE','MSIE','MSIE','Firefox','Safari','MSIE','MSIE','MSIE','MSIE')
par2 = 'grey'
mytab <- sort(table(x),T)
barplot(mytab, col = par2, main = "Frequency Plot", xlab = "", ylab = 'Absolute Frequency')
Figure 55.1: Frequency Plot (based on barplot)

The above code snippet produces the ordinary Frequency Plot based on the barplot function in R (only the height of each bar is meaningful). There are other alternatives based on the (ordinary) plot and dotchart functions. The dot chart variant is sometimes called the Cleveland dot plot (Cleveland and McGill 1985).

Code 
plot(mytab, ylab = "Absolute Frequency", main = "Frequency Plot", xlab = "")
Figure 55.2: Frequency Plot (based on plot)
Code 
taborder <- mytab[order(mytab)]
dotchart(taborder, pch=19, main="Cleveland Dot Plot", xlab="Absolute Frequency", ylab="")
Figure 55.3: Frequency Plot (based on dotchart)

To create a Frequency Plot, the R code first computes a frequency table with the table() function (there is no external library required). The result of this function is used as an input in barplot, plot, or dotchart which renders the plot on screen. The sort function is optional and can be used to order the bars in ascending or descending order. The dataset represents the various browsers that were used by students in 2010 while doing online assignments. At the time, most students used Internet Explorer followed by FireFox, Safari, and Chrome.

55.5 Purpose

The Frequency Plot can be used to graphically examine how often each category occurs in the univariate dataset. The absolute frequencies are ordered from high to low because this allows the user to quickly assess the ranking of each category.

55.6 Pros & Cons

55.6.1 Pros

The Frequency Plot has the following advantages:

  • It is easy to compute.
  • It is relatively easy to interpret.
  • Many readers are familiar with Frequency Plots.

55.6.2 Cons

The Frequency Plot has the following disadvantages:

  • The amount of information that is conveyed is limited.

  • If there are many categories (with relatively low frequencies) the Frequency Plot may be hard to read. Excluding the categories with extremely low frequencies may be necessary to solve this problem.

55.7 Example

The following example shows the status of checking accounts from clients of a local credit company. Note that the categories include both balance ranges (“Negative”, “0 - 200”, “Over 200”) and a fundamentally different category (“No checking account”) representing the absence of an account. This mixture is common in real-world data and illustrates why frequency plots are appropriate: the data is categorical regardless of whether some categories happen to describe numerical ranges.

The interactive R module uses the plot and dotchart functions whereas the old module (which is available on https://compute.wessa.net/histo.wasp) relies on barplot.

Interactive Shiny app (click to load).
Open in new tab

55.8 Task

Compute the Frequency Plot for all operating systems and explain your findings. Describe the format that is used in RFC to generate this plot (for quantitative data). Hint: Chapter 54 explains where the data can be found.