80 types of charts for data visualization. Part 5: Distribution Charts
by Ivan Kilin — Data Visualization Specialist & Dieuwertje van Dijk — Data Visualization Designer @
Recently we’ve rewritten our essential article about 80 types of charts. Our main goal was to make the information structured and helpful right away. We talk about every chart starting from the look and feel and then going into practical advice on when you should and when you shouldn’t use this particular chart type. We are publishing parts of this article here. One chart group per each article.
In the fifth part, we’ll talk about distribution graphs. Distribution graphs are used to show how the data is spread. It is also can be used to show probability distribution. The essence of distribution graphs is to show how often each value appears in the dataset. The shape of distribution helps to gasp patterns which makes them a powerful tool for analysis in many fields. Histogram is the first option among distribution graphs.
Let’s explore distribution graphs one by one. Keep on reading!
Density plot
Alternative names: Kernel density plot, Density trace graph
What is a density plot?
A density plot is a type of chart that helps us visualize how the numeric data is being distributed over a period of time. Density plots somewhat resemble smooth peaks and valleys plotted between two axes. These correspond to a higher or lower concentration of values. A density plot is a variation of a histogram. However, it is visually more appealing, as it loses the sharp edges typical for histograms and adds a smooth continuous curve. The area below the line is usually transparent, which allows to show multiple distributions in one chart.
When should you use a density plot?
A density plot is a good choice for visualizing the distribution of continuous data. The specific case it is most useful is when you have multiple distribution curves and you want to compare them. The transparent nature of density plots allows one to see all the shapes of distribution curves as well as overlapping areas,
When you shouldn’t use a density plot
Avoid a density plot when you need every detail of the distribution. Density plots tend to smooth the details. Also, watch carefully for the readability. Too many distribution lines can make the chart cluttered.
Ridgeline plot
Alternative names: Joy plot, Joyplot
What is a ridgeline plot?
A ridgeline plot is a somewhat special type of chart. A ridgeline plot shows the distribution of a numeric value for several groups of a category. It is done by illustrating partially overlapping line plots (that can be made of density plots or histograms), which then can resemble a mountain range.
When should you use a ridgeline plot?
A ridgeline plot is used to visualize distribution over time (and space). It can be useful in case there are too many distribution lines for a density plot. The structure of the chart allows for the placement of a lot of distribution lines without losing readability. The limit is only the size of the chart.
When you shouldn’t use a ridgeline plot
The form of the ridgeline has some special features. Separate lines should have more or less the same form or have minor step-by-step changes. If the form of the lines is drastically different, some of the details might not be visible as they will be hidden by other charts. So it’s a trial-and-error process for this chart. You should try your data beforehand and see if everything important can be seen on the chart.
Horizon chart
What is a horizon chart?
A horizon chart is what you get when you compress a line or area chart into narrow stripes. To understand how it’s made, imagine you have an area chart with a range from -10 to +10 made of paper. You cut the chart say every 5 notches. After that, you place them above each other in ascending order. The result will be a horizon chart.
The difference with other charts is that both the positive and negative values are shown above the baseline, instead of showing negative values under the baseline. This allows you to show a lot of data in a very condensed manner.
Usually, to create a horizon chart, a diverging color scheme is used. This allows you to show both positive and negative values.
When should you use a horizon chart?
When you are dealing with a lot of categories and you want to make efficient use of space, this chart is the way to go. It is perfect to show time series data on the horizontal axis. Also, values on the vertical axis can be represented by colored bands. This makes it possible to show great precision of the values.
When you shouldn’t use a horizon chart
If you have enough space for a regular area or line chart and your audience is not familiar with the horizon chart, you definitely shouldn’t use the horizon chart. This chart is really hard to read intuitively; you really need to know the structure of how it’s made to use it effectively.
Histogram
Alternative names: Frequency distribution graph, Frequency distribution chart
What is a histogram?
A histogram is a type of chart that visually resembles a column chart. It’s a graph that consists of vertical rectangles (columns), whose length is proportional to the frequency of a variable (data items). The main visual difference between a histogram and a column chart is that there is no gap between the columns. That’s because, in a histogram, the values are grouped into ranges (bins). Each column’s height shows how many items fall within that range.
When should you use a histogram?
A histogram is a first option for visualizing a distribution. They are pretty versatile and can be used both to show an overall view when large bins are used and a detailed view if small bins are used. There’s also a specific formula to calculate the size of a bin for general use.
When you shouldn’t use a histogram
Most of the issues are connected to the visual similarities of the histogram and the bar chart. To recap it once again. Bar charts are used for comparing categories. Histograms are for showing distribution, which shows how many items fall into a certain data range.
Radial histogram
Alternative names: Angular histogram, Circular histogram, Polar histogram
What is a radial histogram?
A radial histogram is simply a variation of a histogram (see above), but with columns wrapped around a circle. It functions the same way as a regular histogram.
When should you use a radial histogram?
If a regular histogram might seem boring for your audience, you can try a radial histogram. It is not as accurate and easy to perceive as the regular histogram, but it’s very likely going to grab your readers’ attention.
When you shouldn’t use a radial histogram
When you need to deliver accurate data to a scientist or analyst, a radial histogram is not a good choice. The more important and serious the data is, the less likely a radial histogram is a good option.
Strip plot
Alternative names: Individual value plot, Single-axis scatter plot
What is a strip plot?
A strip plot is a variation of a scatter plot, but it only has one categorical and one numerical axis. It is a chart used to illustrate the distribution of many individual one-dimensional values. These values look like transparent dots located along a single (category) axis. If some of the dots have the same value, they can overlap, creating something that looks like a strip. Unlike a histogram and a density plot, the values are not placed in bins. Distribution is shown here via the density of overlapping dots.
When should you use a strip plot?
A strip plot works well for showing the distribution of several categories. It works best when the number of values is under 20–30 and there’s no intense overlapping of dots. A strip plot is also useful when you want to show the exact values. Another advantage of this plot is its space efficiency.
When you shouldn’t use a strip plot
When the number of values exceeds 30, the overlapping of dots might look overwhelming and not informative, even if they are transparent.
Jitter plot
Alternative names: Jittered strip plot, Jittered individual value plot
What is a jitter plot?
A jitter plot is an alternative to a strip plot (see above). It is used to visualize the distribution of several categorical variables. The main difference from a strip plot is that the dots used in the charts are shifted on the horizontal y-axis. This avoids overlapping (overplotting), which in turn avoids a lack of clarity.
When should you use a jitter plot?
Jitter plots have basically the same advantages as strip plots, but the number of values that can be shown under one category is more than for strip plots. This is due to additional space that is available via random shifts of dots.
When you shouldn’t use a jitter plot
In some cases, the randomness of the dots shift can confuse the reader. The downside of the jitter plot is that you need to manually control the level of the jitter so that all the dots are more or less visible. Another thing you need to control is the overlap of the dots that belong to neighbor categories; they should be distinguishable from each other.
One dimensional heatmap
What is a one-dimensional heatmap?
A one-dimensional heatmap is a chart that shows the distribution of a variable or category using colored stripes. Basically, you can get a one-dimensional heatmap if you take only one category from a heatmap. The visual difference is that in most cases, narrow and tall stripes are used instead of rectangles.
When should you use a one-dimensional heatmap?
If you want to zoom in on one category and focus on the evolution of that variable, you can use heatmaps in only one dimension. These charts are very popular in climate communication and often visualize temperatures.
When you shouldn’t use a one-dimensional heatmap
As one-dimensional heatmaps depend heavily on color, they can fail to deliver information to people with different types of color blindness.
Beeswarm chart
Alternative name: Swarm plot
What is a beeswarm chart?
In a beeswarm chart, dots are used to show the distribution of a variable. Values are each represented by one dot, and the swarm of dots represents the distribution found in the data. Instead of packing them in bins, the dots are scattered around each other and plotted on a single axis. This kind of chart is very useful when you want to display a lot of data points at once.
When should you use a beeswarm chart?
If you want to use dots to show the distribution of a variable and it’s important that dots won’t overlap, a beeswarm chart is your choice. The additional advantage of the beeswarm chart is that you can assign an additional variable to the size of the dot.
When you shouldn’t use a beeswarm chart
Beeswarm charts are constructed in a way that the dots are placed with some approximation. It means that the position of every dot might not be accurate. It means if the position of every dot is important to you, you should choose another chart: a histogram, a density plot or a strip plot.
Box chart
Alternative names: Box plot, Boxplot, Box-and-whisker plot/chart, Whisker plot
What is a box plot?
A box chart is used to depict the distributions of one or more groups of numeric data. It consists of several elements. The box showing the middle 50% оf the data is also called the interquartile range (IQR). The line inside the box shows the median of the data. The rest of the data is usually shown using whiskers on the left and right sides of the box (in case of a horizontal box plot). Sometimes there are outliers in the data. In this case, the whisker is limited to 1.5 of the interquartile range, and the rest of the values are shown using dots.
When should you use a box plot?
Box plots are meant to provide a high level of information at a glance — a summary of data. This type of graph is quite popular in the research and financial fields. So a box plot is a perfect solution for a quick overview of the data when your audience is familiar with statistical charts.
When you shouldn’t use a box plot
A box plot can be quite confusing for the general public. If your readers aren’t familiar with statistical charts, it’s better to use something simpler, like a histogram or range plot. Also, if you want to show detailed information, a box plot is not the best choice, as it works better for an overview of the data.
Violin plot
What is a violin plot?
A violin plot is a hybrid of a box plot and a density plot. Thanks to this, a violin plot depicts distributions of numeric data for one or more groups using density curves. Visually, it resembles a violin, hence its name.
When should you use a violin plot?
A violin plot combines the best aspects of a box plot and a density plot. This versatility allows it to be used for various audiences, from a general audience to scientists and analysts.
When you shouldn’t use a violin plot
For a small dataset, a violin plot can be an excessive option. It can also fail for visualizing outliers. In this case, a box plot would be a better option.
That’s it for distribution charts. In the next part we’ll explore geospatial maps and other unusual types of charts. Take care and see you soon!
You can also check all other parts here:
Part 1: Comparison charts
Part 2: Correlation charts
Part 3: Part-to-whole charts
Part 4: Data-over-time charts
Part 6: Geospatial and other charts
