I really liked Lin’s presentation that she gave to our class. I think she did a good job at explaining the importance graph’s have in the digital humanities. I must admit I am a little bias as she is my boss (I work in the Quantitative Resource Center at Carleton College). I also think she did a great job at emphasizing that not all graphs work for all variables. For this weeks lab assignment I used Rawgraphs.io to make graphs to capture a dataset on the most popular baby names in New Zealand from 2001-2010. I focused on trying to capture as many of the given variables as I could while still making the graphs easy to read.
This is my first graph which is a line graph. I think it does the best job at capturing the dataset but it does appear to be too crowded!
This is the graph that I decided did the best job of capturing the whole entire picture. Below are some more digestible graphs that I also made. To make the line graph above at all legible, I had to select a different color for each name. For default, about half the names were given the same grey color. In addition, I also had to adjust the width and height so that the names were not stacked up on each other. At first, this graph was much more clustered.
This is my pie chart graph. I think this is the cleanest design although it lacks the split up of the years. I think it does a the best job of conveying the bigger picture.
After making these two graphs I decided I wanted a graph that showed all the information but was easier to read. I kept in mind Lin’s advice and tried to make sure I as picking a graph that fit my variables. I decided to use a graph that showed the top names for each year on a separate graph. I had to make the width and the length of the graphs longer so that the names were readable and not stacked on one another. Although I still feel as though this was not as visually appealing as my other attempts.
This is a graph I made with each graph representing one year.
In all, I feel as though none of my graphs were perfect, although I think I did a good job of exploring the different ways this data could be represented. Reflecting, I kind of wish I could have just uploaded this csv into R and played around with it myself. I really like using R. I think I would have eliminated a lot of rows as I feel like looking at half of the names would have told the same story. Regardless, I feel as though by looking at all three of my graphs, a person could get a really good idea of the most popular baby names in New Zealand from 2001-2010.
2 thoughts on “Clay’s Graphs!”
Leave a Reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
I like the variation you had in graphic display types! I think you’re absolutely right with the Pie Chart issue—the ability to make arcs based on year just didn’t seem to work properly; only creating one or maybe two arcs which were still unlabeled. I also experimented with the line graph, I think that one would be the most useful for observing name prominence over time, but the data labels not being modifiable made it harder. Well done!
I really appreciate your concluding paragraph. I too found it very difficult to make or find the “perfect” graph type for what I wanted to display. It definitely made me understand how every type of graph has both its strengths and weaknesses regarding complexity level. I wish we were tasked with editing the data itself in a program like R too. I was struggling to remove duplicate names (i.e. Jack being mentioned numerous times over the years) and wish we could’ve used R to fix that problem.