Lab Week 3: Data Visualization

Given the large amount of data in the given file, I chose to use Flourish to create an animated visualization that would highlight only certain pieces of information at a given moment. More specifically, I chose to make a scatter plot in which only the top ten girl/boy names are present, rather than having all possible options lining the x-axis. There is a play button at the top of the graph that can either progress automatically or be manually scrubbed, and which shows the movement, addition, and removal of the most popular baby names. This prevents visual overload by only showing twenty names at a time, rather than the thirty-three total options, and allows users to more clearly see which names are coming in and out of style as time progresses. The names are ordered by gender as well as by rank, the latter of which corresponds to the number of babies born with that particular name. As such, the female names are on the left half of the graph, moving from left to right from most popular to least. The same is true on the right half of the graph with boy names. I changed the graph settings so that each data point is sized according to the count of names, where higher counts correspond to larger points; within each gender, the points are color coded so that the most saturated point has the highest count. These design choices are meant to provide visual contrast and distinction between the different data points, as Lin discussed in the lecture, as well as to provide an engaging demonstration of trends over time. 

The choice of a scatter plot is appropriate for the given data as it shows the count of each name in relation to each other, without necessarily being visually overwhelming or intelligible. This was also aided in large part by the use of animation for the temporal aspect. At first, I had experimented with using visualizations that would track popularity over time, such as line charts and violin plots. However, given that thirty-three names were present as data points, displaying the name and count for ten years at once proved cluttered and difficult to discern meaningful patterns or distinctions between data points. The elimination of lines in the scatter plot prevented overlap between the points, and the addition of animation added the missing element of visual cleanliness that it was missing before. 

I found Lin’s lecture on the creation of data visualizations and the readings on their role in DH to be very illuminating. While my first instinct in this lab was to immediately choose a graph style, I quickly realized that this would cause more frustration than it was worth without first looking through the data file and noting patterns on my own. This period of exploration, as Lin called it, allowed me to notice trends in the changing popularity of names that I was then able to reflect in my visualization (the top boys’ names stay much more consistent then the girls’ names, for example). I was also forced to change my approach when certain qualitative fields weren’t accepted as values for the graph, which calls back to humanists’ concerns about the use of subjective and qualitative information as data. While it was fairly easy to work with data like names, I see how it could become challenging to input more complex qualitative data into quantitative visualizations like graphs. A visualization such as the one used here could be useful in DH projects to map the use of certain names or words over time, which could be utilized for English or history scholars. The approach benefits from a smaller sample size – something like ten-twenty values – which could be more in line with the scale of certain humanistic inquiries. On the whole, the use of data visualization seems like it could have a lot of uses for DH, though the process of arriving at a useful visualization would perhaps take more trial and error than in a STEM field.