<TODO: Write a compelling but concise header here>
I Hate Statistics Interviews: Tariq Khokhar! At the United Nations World Data Forum, I sat down with Tariq Khokhar from the World Bank. I asked him about his experiences summarising data. His answer: focus on the headline! People remember the headline twice as often as the graphs.
How it started: Nobody is reading your PDF
Pim: “Tariq, thank you for taking the time to sit down with me here while we are at the United Nations World Data Forum. Could you briefly describe how and why you started creating charts at the World Bank?”
Tariq: “Sure. A few years ago I came across an article in the Washington Post with an interesting but also quite alarming message: Nobody is reading your pdf. It reported one of our own investigations that had found that some reports were just never being downloaded. Working as a data scientist at the World Bank myself, this really grabbed my attention of course.
Now it wasn’t that these reports did not contain interesting information. They definitely did. The problem was that the interesting information in these reports was being buried so deep, that nobody looked at it.
So what I started to do was to take these big reports and try to summarise them in a graph. I shared these on my blog.
An example is the number of people that do not have access to adequate sanitation. People read this headline and suddenly they are interested: 2.4 billion people that do not have access to a toilet? That’s a lot! Then you ask: where do these people live? It turns out that this is something you can visualise neatly in a chart.
Scaling: teaching communications people to use graphs
My blogging work kept on growing, until finally it took up more time than my data science work. So I decided to train the people from the communications department how to do this as well.”
Pim: How did they learn to create good visualisations?
Tariq: “I think communications people actually can be very good at this. They have been trained to be very careful with words. The only thing I did, is train them to be just as careful with data and visuals as they are with words.”
People recollect the headline 2x as often as the visual
Pim: What makes a good visualisation?
Tariq: “I tend to take my inspiration from scientific research on how people perceive visualisations. Of course, this is a broad field and there will be papers describing different findings. But one paper that I found very interesting (Borkin et al) is one where they used eye tracking and recollection surveys to investigate what people take away from looking at a visual.
The findings were counter-intuitive: it turned out people spent most time reading words. Summary: People read words.
People also recollected the headline 2x as often as the information in the visual. So if you do one thing: think deeply about the headline. Make it compelling but clear. Do not write “relation between income and health.” But: “rich people live longer” (or: long living people are richer.. 😉 correlation/causation..)”
Tariq: “the second piece of advice would be: confine yourself. Tools can help here. The visualisation software that we use restricts us. We can — and always have to — write a headline, subtitle, chart and one paragraph that includes the source. That’s it. That’s all you have.”
Try it yourself
Can you create a good summary of a report or research that more people should know about? Try it yourself: write a compelling, concise headline, a subtitle, one chart and one paragraph that includes the source. Please share your results with us! twitter.com/ihatestatistics #nobodyreadsyourpdf
Iterate and see what works for your audience
Pim: there are two things that I personally take away from my interview with Tariq.
First: I did not expect the headline to be so important. This interviewed showed me how important the choice of words in the headline is.
Secondly: what I really like about Tariq’s approach is how he tailors his approach to his audience. He doesn’t do things just because that’s how they have always been done. He uses findings from scientific research to guide his work. Then he designs, he iterates and he finds what works to communicate the important insights the World Bank has to offer.
This is also what we at I Hate Statistics strive towards. Based on scientific research, design principles and a lot of iteration, discover how we can best help people become learn to appreciate and understand data and statistics.
So thank you very much Tariq for sharing your experiences and lessons with us!