Examples of Datawrapper charts that show complex data
This is a collection of charts that show uncertainty and percentiles, lots of data points and a rolling average.
If you want to find out how a certain chart was created, hover over the chart and click on "Edit this chart" in the top right corner. This will create a copy of the chart and bring you right into the step 3: Visualize of the chart creation. You can play around with the settings or go back to step 2: Check & Describe, to find out how the data for this chart was structured.
Percentiles and outliers
Percentiles and outliers give us information about how our measure points distribute. According to Wikipedia, "a percentile is a measure used in statistics indicating the value below which a given percentage of observations in a group of observations falls. For example, the 20th percentile is the value (or score) below which 20% of the observations may be found. Equivalently, 80% of the observations are found above the 20th percentile."
Here's a concrete example: If you have a group of nine people who are ages 2, 5, 5, 5, 20, 60, 90, 90, and 93, then the group's mean age is 41 years and its median age is 20. But you might want to know the median age of just the younger half of the group. That would be the 25th percentile (5 years). Or you might want to show the 75th percentile, which is the median age of the older half (90 years).
And you have extreme outliers (2 and 93) that you might want to mention in your chart, too.
If you're coming from a scientific background, you know that a good chart to show percentiles is the box plot (also called box-and-whisker plot). Since this is a chart that's hard to understand for the mainstream audience, we don't offer it. However, there's still a way to show percentiles and outliers (minimum/maximum) with Datawrapper. In the following chart, we used a dot plot:
Read more about the chart in the chartable blog "The black dots that warm our planet".
Uncertainty, estimates & forecasts
The data you show is often the best guess instead of the ultimate truth. In the best case, your data source provides numbers that show how good that guess actually is, like confidence intervals.
Here are three charts that show uncertainty, two line charts and one stacked bar chart.
Read more about the chart in this chartable blog "The only chart we should be looking at".
Read more about the chart in this academy article "How to show confidence intervals in Datawrapper line charts".
Read more about the chart in this chartable blog "Weekly Chart Depth and breadth in charts".
Show lots of measured values
To increase the nuance in your data, consider showing not just one line, but lots of them. A scatterplot is a great chart type to do so. The following chart shows a line chart. The trick here is to set the line width to 0px while keeping the line symbols:
You can read more about the chart in this chartable blog "Greenland’s ice is melting, but without an OMG moment".
If you have fluctuating values and you want to smoothen them out, a rolling average (also called moving average) can be a good idea. Similar to the chart above, showing both the measured values and the average can be done by setting the line width to 0px:
But there's another way: Create a "base" that stays at zero, then fill the area between the measured values and this base line with a light grey color and set the line width of the measured values to 0px again:
You can even set different interpolations to the filled areas so it looks like a chart that is a combination of line and column charts. To read more about this chart, read the weekly chart blog " The COVID-19 chart I wish I didn’t have to make".