Doing the 

Bar Charts Right

In order to let you create good quality charts we have built a number of features and settings into Datawrapper to achieve just hat. 

This post provides an overview on how we design Datawrapper to assist in creating high quality. Hint: No as easy as it looks. 

Datawrapper is a browser-based tool, which means that we work hard on keeping the balance between simplicity and complexity. The goal is to make Datawrapper the best viable option to create good charts in very short time and help to avoid the typical and most common charting mistakes. 

Make no mistake: Excel is great, Illustrator is great. But Datawrapper is actually about achieving something else: Letting editors create high quality charts quickly, without compromising the quality and visual clarity. So far we are very successful doing just that, with more than 250.000 charts created by more more than 40.000 users. 

What makes a good chart?

A good chart tells a story "simply, clearly, and accurately" (Stephen Few, Perceptual Edge, 2007). This is not easy and requires some experience. To get to this goal there are some generally accepted rules (some say recommendations) to achieve this. 

Rules for good charts are based on  how the human eye will detect the pattern we show. We all have this amazing ability to see and understand a good "picture" in fractions of a second, the role of a chart is to present data as a pattern that shows a development, a change, a comparison or a distribution. 

Data-ink ratio

The most important insight is that a chart gets better when you leave out decoration. In short: To get to a good chart you want to check several times what to leave out. 

The rule to follow is to achieve a good or low "data-ink ratio". The most "ink" in a chart should be used for the data, not for supporting elements. This rule has been defined by Edward Tufte in his book "The Visual Display of Quantitative Information (1983)".

Less "ink" means that the data can speak for itself, whether in a line chart, a column chart or a map. Forget about multiple colors just because they "look nice". Stay away from 3D bar charts, even though the charting programm you used so far offers them. They are just harder to read.

Datawrapper Bar Charts

What other rules are there?

Based on another great book about good and bad charting practices, here are the main things to know (including live examples from Datawrapper). 

"The width of the bars should be about twice the width of the space between the bars" (Dona M. Wong, The Wall Street Journal Guide to Information Graphics, 2010). 

This simply helps to give the bars room and helps the user to see the pattern. The columns or bars should be flat, use the most minimal color possible (e.g. no shades, no grids, lines around the rectangles not too thick). 

"A lighter shade can be used to distinguish projections and estimates from actual values". (Dona M. Wong)

Start at the zero baseline - No exceptions

For bar and column charts this is a hard rule. If you don't show the zero baseline this truncation "obscures the discrete  will join the ranks with some media outlets which are ridiculed for their bad charting practices. 


Other Interesting Tutorials

How to create a Stacked Column Chart

Learn how to create a stacked column chart in minutes with Datawrapper.

View Guide

All-new line charts

Datawrapper has completely revamped line charts: More responsive, cleaner and with annotations.

View Guide