- Megan Johanson
Same Data, Four Visuals: Midterm Voter Turnout
There is no one right way to visualize a set of data. Often, the type of chart you choose will depend on the story you are trying to tell because different charts serve different purposes.
I recently came across a table created by NBC to show changes in midterm voter turnout between 2014 and 2018. The data is from the U.S. Census Bureau and NBC broke down the voting rates into four age groups.
I thought it would be a good opportunity to compare the impact of using different chart types to visualize the same data. To me, the most interesting part of the data is that the largest increase in voter turnout was in the 18-29 age group, so that is what I tried to highlight.
Typical Table
Here is the table from NBC, with a link to the original article.
Table Assessment
The yellow boxes highlight the age groups with the largest change, but you have to spend a lot of time comparing all the percentages to know how big the increases were and how the turnout differed by age group. The table also leaves the reader to determine the takeaway message on their own.
To test out different chart types, I recreated the table in Excel and then thought through which types of graphs could be used to depict the data. A bar chart, slope graph, and dot plot would all work but it's hard to know which would tell the story the best without seeing them filled out with the real data.
Basic Bar Chart
First, I created a basic horizontal bar chart in Excel. This is what it looks like with Excel’s standard formatting.
Below is what the bar chart looks like when I applied data visualization techniques designed to tell a story. I strategically used color, directly labelled the bars, removed unnecessary lines and numbers, and used a chart title that clearly states the main point.
Bar Chart Assessment
Bar charts are nice because everyone knows how to read them and people are good at judging differences in the lengths of bars. However, one downside is that the bars are very visually heavy, and the eye is drawn to the longest bars. In our case, the longest bars are not the most interesting part of the data set.
Sleek Slopegraph
To create a slopegraph, I started with the same table and inserted a scatter plot with smooth lines and markers. This is what the default from Excel looks like.
Next, I switched the row and column under Chart Tools-Design. Then, I adjusted the x-axis so that it only showed the years 2014 and 2018 and made the y-axis to go up to 100%. Once it was scaled correctly I deleted the numbers. I also removed the chart boarder, y-axis grid-lines, x-axis line and title, which I’ll replace later.
This is what the chart looks like now:
For the finishing touches I moved the two lines closer to each other, changed the colors and font, removed the legend, and directly labelled the lines. Finally, I added a title that clearly states the most interesting point of the chart. You can find a more detailed tutorial here.
Slopegraph Assessment
The slopegraph enables me to highlight the change between years for one age group rather than color coding the two years in the bar chart. It is not as visually heavy as the bar chart and allows readers to quickly understand the amount of change. In addition, it makes it immediately obvious that even though the 18-29 year olds had the largest increase in voter turnout, they had the lowest turnout rates of any age group both years.
Distinctive Dot Plot
Next, I created a dot plot. This is a pretty complex process, but there are step-by-step instructions here you can follow. Let’s just say it requires a fair bit of manual editing to achieve but the end product is worth it. You need to modify the data table to space out the data and then insert a scatter plot without lines.
And here is what the dot plot looks like after additional formatting to increase usability. As with the other visualizations, I strategically color coded the data, added direct labels, and made the title informative.
