Exploring + ‘visceralizing’ an incomplete dataset
In DESIGN 384: INFORMATION VISUALIZATION, a course I took at the UW, students were asked to visualize a database of accession data from the Seattle Arboretum.
Directly south from the University of Washington’s Seattle campus is a massive, 230-acre green space. It sits in the heart of Seattle, marked by bike paths and running trails which snake through its terrain.
It is a total gem, and importantly, it contains a dynamic assortment of over 40,000 plants—some found nowhere else in the Northwest—cared for by volunteers, arborists, and gardeners.
Since 1936, over 12,480 individual plants have been accessioned and digitized into the Arboretum’s database. These accessions can be categorized into 139 unique families + 3982 unique species (many of which are classified as threatened or endangered).
For example, in the dataset, you can find that 491 individual trees belonging to the Cypress family have been accessioned across the Arboretum.
The goal for this project was to set up a physical installation in the Arboretum’s visitor center, and present our project to the public.
I created a 4-dimensional bar chart, which represents the number of accessions + unique species for each cell of the Arboretum’s grid.
“The idea was to try and capture plant diversity.”
Constructed from laser-cut plywood, each square block represents the presence of five unique species (to its corresponding location in the Arboretum). Inside each block is a wooden rod, whose height represents the number of individual accessions.
The idea was to try and capture plant diversity. Because the heights of blocks + rods are proportional, a viewer can immediately see that some areas of the Arboretum are both more dense, and more diverse.
Prototyping with D3.js
While trying to wrap my head around the dataset and find stories within it, I found I needed a visual representation of it... and pivot tables only went so far. I decided it was time to learn a new skill, and installed D3.js, a data visualization library for the web.
This proved to be a very fruitful decision. It took some trial and effort to set it up properly, but being able to explore the dataset in terms of its placement in physical space gave me a much clearer picture of what I was working with.
