Exploratory Data Visualization

Originally I did make a line graph, but I felt like there was a better way to express the data, so I opted for a bump chart. Only in the past week have I realized how many different ways of expressing data there are; the number of possible types of graphs and visualizations are slightly overwhelming. The dataset was a small csv containing the top ten names for boys and girls (including their count) in New Zealand from 2001 to 2010. After reading the data we were to display, I knew needed to visualize progression in time that would allow for comparisons between the popularity of different names in a readable way.

A bump chart is similar to a line graph, but the thickness of the lines indicate a chosen quantity. In my case this was the sum of each name that year. This aspect of a bump chart is useful because we can observe constant change in line weight and can track the relative popularity of a name. For instance, the name Jack was wildly based on its line’s weight and position, but it took a rapid downturn after 2009. Unsurprisingly, Liam, my name, rose to take its place in the number one spot.

In addition, I adjusted line position to reflect the ranking of name popularity between years. If a line is above another, that name was ranked higher (or more popular) during that year. For instance, the two names, Joshua and Jack, were on the top of the graph for the majority of the early 2000s, so can infer they were the most popular names for a boy during that range.

I also separated the data by the sex associated with the name; the data would have been incomprehensible otherwise. After spending time with the data, this distinction in the data led me to (a rather delayed) reflection on our reading, What Gets Counted by Catherine D’Ignazio and Lauren Klein. By making this decision to categorize and separate the data, we are ignoring intersex and non-binary people for the sake of readability.

The names themselves appear inline with their line in order to add to readability; the software (Rawgraphs.io) does not allow much customization in this area, and I felt this was far better than the alternative as you can identify what name corresponds with each line. Alternatively, there was the option to add a legend, but it was not large enough to hold all the terms as I was not able create separate legends for boys and girls.

Furthermore, the main lesson I pulled from Lin’s lecture was guiding your audience towards what you want them to see or understand. She described you can draw your audience’s eyes towards certain trends in the data (like a linear trend in a scatter plot) using color and lack there of. There is no one point or name I wish to draw my audiences eyes towards. Instead, I want them to consider the overall changes in popularity between all names. I used different colors to aid comprehensibility, so an observer would not confuse lines with one another.

-Liam Keane 1/28