Gallery

Gallery

The gallery should help you to see the themes and colors applied in different plots. This template can also serve as an inspiration to get an idea of which graphics might be suitable for your case. Note that this template should not be considered as a 1 to 1 guideline but rather as a tool which helps statworx to create a more coherent picture when visualizing data.

Distributions

Density chart

Density chart multiple

NOTE: For more than 3 densities the plot can get very cluttered. While you change the plot type to violin plots or jittered box plots there are further alternative ways of displaying densities as depicted in the following tabset which help to reduce the clutter.

Alternatives denisty

No Fill

Facet

Violin chart

Ridgeline

Histogram

NOTE: When making a histogram, always explore multiple bin widths. In addition, avoid overlapping histograms as they can be confused with stacked bar charts.

Boxplot

NOTE: Adding jittered points to a box plot is useful to see the underlying distribution of the data.

Trend & Time

Series

NOTE: For a single line one can also fill the area under the curve with a solid color. This choice can emphasize a trend in the data, because it visually separates the area above the curve from the area below.

Series multiple

NOTE: For multiple time series there is danger of visualizing a spaghetti chart. Hence, one technique to reduce the mental load is to directly label the lines since matching multiple lines with the legend is cumbersome. Alternatively, if the clutter is too overwhelming the following tabset presents further techniques to plot multiple series.

Alternatives multiple series

Facet

Stacked Area chart

NOTE: Consider this type of chart if you attempt to visualize an evolution of the whole and the relative proportions of each group. However, use stacked are charts cautiously since they can be hard to read due to moving baselines.

Forecast

Relationship

Scatter chart

NOTE: The second plot shows a common problem to scatter plots, namely overplottling. The following tabset of alternatives present some easy workarounds. If a categorical variable is present common techniques such as facetting or highlighting apply.

Alternatives scatter

Dot Size

Transparency

Bubble chart

NOTE: The differences between values encoded as bubble size are harder to perceive than differences between values encoded as position. Hence, the “All-against-all chart” is on alternative. In addition, one can also opt for a correlogram as described in the next section. Besides, bubble charts can also suffer from overplotting.

All-against-all

Correlation

NOTE: When we have more than three to four quantiative variables, all-against-all scatter plot matrices quickly become unwieldy. In this case, it is more useful to quantify the amount of association between pairs of variables and visualize this quantity rather than the raw data.

Comparison

Bar chart

NOTE: One should only rearrange bars, when there is no natural ordering to the categories. Whenever there is a natural ordering (i.e. when our categorical variable is an ordered factor) one should keep the original ordering in the visualization.
When bars are of similar length it is visually less appealing to use bar plots (“Moire effect”). In this case, one can resort to use Lollipop charts.

Lollipop chart

Paired and stacked bar chart

NOTE: For multiple groups and subgroups the comparison gets more difficult across and within. the bars. In addition, the clutter of the plot increases. Two alternatives to the bar chart are presented below.

Dot chart

Waffle chart

Marimekko chart

Heatmap

Part to whole

100% Stacked Bar

NOTE: The barplot is the best alternative to pie plots. However, the example below shows that for similar shares it is still difficult to compare them within and across the categories. So for this example, side by side bar charts are a better choice to visualize proportions.

Side by side bars

Side by side lollipop

Waffle Chart

Tree Map