A picture may be worth a thousand words, but the right data visualization can be priceless.
Data visualization has become a creative and cutting edge method of portraying information for storytellers, journalists and businesses. From the simplest of graphs to marketing infographics to complex data journalism, data must be presented accurately, not just in the way that’s the prettiest or best supports your point of view.
Choosing the Wrong Visualization
Choosing the right way to showcase your data is essential. Think about what you’re trying to convey: a comparison? A relationship? A distribution? Here are a few basic guidelines you can go by:
- Use a table to show qualitative information or compare pairs of related values
- Use a line chart to show changes or trends over time, or relationships within a continuous set of data
- Use a Venn diagram to compare similarities and differences between two or three items
- Use a pie chart to show parts of a whole
- Use a bar or column chart to compare different items
- Use an area chart to show trends over time (like a line chart) and to show volume (in the space between the axes and line that’s filled in)
The website Which Graph can help you determine which data visualization is the right choice.
This should be obvious, but using bad data will mess up your data visualization. One example of this is combining data from disparate sources to tell the story you want told, and not what the data actually says. For example, in 2015, the White House tweeted this chart:
This chart cites the National Center for Education Statistics at the U.S. Department of Education as the data source, but the NCES doesn’t provide a dataset on graduation rates for 2008 to 2014. The White House must have pulled this information from different colleges, which could have provided graduation rates in different ways.
Keith Collins at Quartz improved this chart by using a dataset from a single source and providing much more context:
As you can see, graduation rates were already picking up Pre-Obama, but the rate is the highest in 2012.
Illustrating elements on a chart can get tricky, as you can see in the White House chart. Does a certain number of books equal a certain percentage? There is no meaning to it. Illustrations can be eye-catching, but make it difficult to actually understand the data. Take this example, from the National Center for Complementary and Integrative Health:
It’s hard to compare the values in this infographic due to the illustration. It would be much easier to interpret the data in a bar chart.
Although illustrations may be visually appealing, they often aren’t the most accurate way of representing the data.
This is something we all covered in elementary school, yet so many data visualizations get this wrong. Having axes with correct labels and standardized axes is critical for representing data correctly. Take this chart:
The X-axes changes from yearly to quarterly, misrepresenting the information.
Charts Within Charts
Placing a chart within another chart is usually a bad idea because it leads to obscuring the information somehow. For example, the chart below is unclear in the information its presenting. I know that it’s trying to show how hospitals are improving, but it’s hard to tell what’s going on here.
Are hospitals improving? This graph from the CQC tells the story pic.twitter.com/7RbTDrtN4H
— Jeremy Hunt (@Jeremy_Hunt) November 2, 2017
Not Adding Up to 100
When making pie charts, it’s essential that the parts add up to the whole – otherwise, the data will be misrepresented. For example, take this infamous pie chart from Fox News:
— Bruce Segal (@besegal) September 30, 2016
Since it looks like those surveyed could choose more than one candidate, the numbers add up to over 100. This data would be better portrayed by a bar chart.
Data Visualization: Be Careful
Whether we realize it or not, we see forms of data visualization all the time, from the nutrient facts table on our food packaging to the graphs shared on the news. Deciding what you want to do with your data and making sure you don’t present too much data in one visualization are both key factors to the clarity of your visualization.