Discourse through Data


On my essay about trees (not the illegal kind)

By Ram Iyer

January 29, 2019

For my model essay, I wrote about the trees of Pittsburgh. Well, that isn't entirely accurate. What I should be saying is that I set out to write about the trees of Pittsburgh, but I ended up down a rabbit hole that ended with me discussing a variety of things. That being said, that variety all had to do with trees in Pittsburgh, so maybe my initial statement was accurate after all.

In order to start this assignment, my first real hurdle was trying to figure out what I wanted to talk about. This was a very interesting problem for me as I started browsing the WPRDC's website, and I found a large amount of interesting datasets that I was curious about. When I went to visualize them in Tableau Public, I couldn't find points that interested me to the point that I wanted to talk about them.

Then I went to Schenley for a bit to clear my mind, as I usually do. I walked around, and noticed a strange assortment of trees. I made a mental note, and then went to do some research after I got home. That's when I decided to search for a dataset on trees on the WPRDC website, and found the dataset that I used to write the model essay (this is it, by the way). I then got to work.

The one issue that I ran into when writing my article was trying to figure out two things:

  1. What to talk about in particular when relating to the trees of Pittsburgh
  2. How to represent the data in such a way that can make it useful to the narrative

After thinking about it for a while, and sleeping on it, I decided on writing the essay in a more personal style. I thought that the easiest way to convey the interesting points about this dataset would be to interact with it along with those who are reading my essay. This allowed the words to flow significantly easier than if I were writing a rigid, formal paper and I thought best reflected my thoughts and findings. 

From there, I wrote a quick introduction documenting my initial thoughts on the paper, and my inspiration for picking the tree dataset. From there, I decided to do some additional work in terms of actually presenting the dataset. At first, I was set on trying to represent the data on a map. The dataset included longitude and latitude coordinates, which was a huge step up from some of the other datasets that I was considering, and it was reasonable easy to filter through important details (how damaged the trees were, the types of trees, etc.). After attempting a few different visualizations with the map, I found that none of them were particularly interesting. There weren't any parts of Pittsburgh that were populated by only a few types of trees, and most of the other measures or dimensions didn't provide any meaningful insight either. 

This led me to pursue different types of visualization, and I eventually decided on a bar graph that compared the population of every tree that had a population of 500 trees or more (I had to reduce the resulting list due to "unknown" or "unrecorded" tree species). This view was the most eye-opening, particularly because of how many foreign trees had a presence in Pittsburgh. This led me down the rabbit hole to investigate whether or not I could get any major insight on why this was.

While the actual search for the origins of these trees didn't yield any fruit, I found out about Pittsburgh's interesting disposition towards the Ginkgo tree, a tree found in east Asia that produces fruit with a pungent smell. This also led me to TreePittsburgh, an organization dedicated to preserving and growing the urban forests around Pittsburgh. With this, I thought that I'd have an interesting story to tell, and it's all thanks to the dataset for pointing me towards it.

Lastly, the aesthetics. I saw Dr. Lavin's post, and believed that a minimalist view would work very well with this kind of article too. I originally messed around with adding pictures of trees and other "interesting" aesthetic choices, but decided against all of them. While I did copy Dr. Lavin's basic aesthetic, I decided to add section titles to help split up this rather fragmented article, and I also implemented superscripts that would redirect to their respective footnotes (granted, most of them are close together at the bottom of the page, so I ended up just creating 5 different links to the bottom of the page). 

All in all, I was led to this dataset from an inspiring moment in Schenley Park, and I'm glad that I had it. I found out about a cool story that I would have never found out otherwise, and I got to write an article about a topic I'm not too familiar with, which was an awesome experience. If I were to develop on this more, I would probably look at which trees provide the most benefit in terms of carbon dioxide on average, and which trees tend to get the most frequently damaged. You can view this model essay on my Github Pages website. Thanks for reading!