Common Data Visualization Mistakes Part 5: Figures Not to Scale

This video is part of the How to Avoid These Data Visualization Mistakes series, presented by Naomi B. Robbins, Data Visualization Expert at NBR.

Transcript:

Figures not to scale. This one I found in my local library. It tells you how many police officers there were in different counties of New Jersey. I'd like you to concentrate on the two that have the red arrows. Passaic County, where I lived at the time, had just over 1000 police officers. Salem had just 100. Does the line for Passaic look 10 times the length of the line for Salem? Why use a graph if it's not proportional to the numbers it's representing? Here it is accurately, and Passaic, about the sixth one down and it does look 10 times Salem, which is the smallest one at the bottom.

This one shows the top 20 tourist-generating countries to the United States, and it's a nice eye-catching figure, but the numbers are difficult to read, so I will read them. If we look at the beginning, it has Canada and Mexico. Canada sends roughly 14 million. Mexico sends just over 4 million, but they look almost the same length, 14 million and 4 million. Let's look at Germany and Japan. One is just over 1 million. The other's just over 2 million, but they look the same. Again, what you see visually has to be proportional to the numbers they represent.

Here's the frequency of eye exams. For people who wear eyeglasses, they tend to be seen once every 1.9 years. Contact lens users tend to be seen 1.17 years. 1.17 looks much bigger than 1.9. It got explained to me, "Well, they're seen more frequently so they're bigger," but that's not what the numbers say.

Here are summer Olympic medals by country. Germany at the bottom has two bars for roughly 500, so if we go to Russia that has 1000, we would expect it to be double. We would expect, if 500 is two bars, 1000 should be four bars. It's five. and then USA has 1975, four times Germany, not three. It's just not to scale at all.

This one came me in the mail requesting money. It was a very worthwhile charity, a charity I respect and gave money to, but I thought this appeal was so deceptive. Notice that it says expenses in red, and we see this little sliver, which gives the visual impression that it's just that little sliver that is expenses. But when you read it, that little sliver says management, management in general, 7%, and that green thing, the next section, which is almost the same color as the yellow, says, development 10%. Well, certainly that red does not look like 7/10ths of the green. And it says the program is 83. Well, as I said, it has development blending in with the program, and the management and development are not to scale. So you get the impression if you give here, it's all going to the program, which is not true at all. It clearly says only 83% goes.