Five Tips for Journalists Doing Data Visualization
Yamkovenko, Stephanie, The Quill
It's not just for the designers and 'Web folks'
General Electric is doing it. Statistical software behemoth SAS is selling software for it. Harvard is teaching courses about it. Data visualization seems to be everywhere.
Journalists need to get involved to provide something that can easily go missing in data visualization: the story. Data visualization should not only describe a dataset, it should also provide context and depth to the data. It should tell a story.
I will share my five tips on how to tell compelling stories with data visualization. Although I wouldn't consider myself an expert on the topic, I recently won The Economist-Nielsen Data Visualization Challenge. (See more on the challenge and winning entry at innocentive.com/NielsenFinalists.)
The challenge focused on analyzing a dataset provided by Nielsen and telling a story using data visualization. After seeing the challenge advertised in The Economist, I decided to partner with my husband, Bogdan Yamkovenko, a researcher and assistant professor with an affinity for statistics.
Having never collaborated on a professional project before (and the fact that neither of us currently works as a graphic designer), we weren't too confident about our prospects for winning. When we got the news that we won and had the opportunity to present our findings at The Economist's World in 2013 Conference in December, I realized there is an important role for journalists in data visualization. Using my experience with the challenge, I developed the following tips for journalists.
1 DONT IMAGINE AN INFOGRAPHIC
The first tip is to stop thinking about an infographie. It can be difficult, because ultimately graphics will be a big part of data visualization, but for the time being focus on the data. It's similar to not writing an article before interviewing sources and doing research.
Bogdan and I began our work on the challenge with a brainstorm about the Nielsen global dataset, which consisted of the consumer confidence index and other data about consumer spending and purchasing habits. Look at the data and see what is there and whether you can see any potential stories that need telling.
2 WHAT'S NOT IN THE DATASET?
Chances are the dataset itself is not complete. Can you supplement the dataset with other data? Think of ways that supplemental data could offer context to the current dataset. Again, journalists rarely write an article after interviewing only one source. Why would that be different with data?
We decided to supplement the Nielsen dataset with the widely available economic indicators such as unemployment rates and the current account balance as a percentage of GDP, which is information on the final pages of every Economist issue. By adding this information, we were able to look at whether the consumers' confidence in the economy matched up with the country's actual economic performance.
3 WHAT IS GOING ON WITH THE OUTLIERS?
Take a look at the dataset that you have supplemented with other relevant data and see whether you find any patterns, outliers or strange connections. After supplementing the Nielsen dataset, we started looking at the outliers (the countries with the highest and lowest consumer confidence). We noticed that countries that had high confidence in the economy were not necessarily the best performing economies. Countries such as Saudi Arabia and Egypt had high confidence, but their economies weren't doing that great.
At this point, we revisited tip 2. We wondered whether democracy was playing a role in the citizens' confidence. We decided to include the Reporters Without Borders Press Freedom Index in our analysis.
4 USE STATISTICS
Yes, many of us went into journalism because we don't like numbers. In graduate school, one of my journalism professors assigned us the textbook "Statistics for People Who (Think They) Hate Statistics." Even if numbers are not your forte, it helps to have a basic understanding of descriptive and inferential statistics; I still revisit that textbook for a refresher on statistics. …