Intro
While learning how to use python to loop through JSON files in my Programming for Cultural Heritage course last spring 2020, I came across a Directory of Playgrounds JSON file from NYC OpenData. The dataset also included borough information and which playgrounds had functioning water fountains. Inspired by my project looping through the dataset last year and by my heightened search for open spaces, parks, and playgrounds to explore during the COVID-19 Pandemic, I wanted to work with visualizing playground and parks data. For this lab, I decided to explore possible correlations between the location of playgrounds and the number of schools within a school district. I also wanted to explore the data to see if there are relationships between the location of NYC Parks and playgrounds/schools.
Inspiration
- The Trust for Public Land (TPL)
TPL was founded in 1972 to champion access and benefit of parks to communities who need them the most. In 2018 the foundation launched ParkServe, a database of parks for 14,000. ParkServe tries to measure who does and does not have access to parks within a 10-minute walk. Though this example is not exactly a map visualization, the website provides a visual breakdown of park/greenspace accessibility based on Age and Race/Ethnicity. Given more time, I’d be really interested in exploring their data collection methodology. Here is a screenshot of their New York City parks ranking.
2. Isamu Noguchi Museum Archives
Another source of inspiration has been my work in the archives at the Isamu Noguchi Museum. Recently, I cataloged a selection of architectural designs for Moerenuma Park in Sapporo, Japan. Looking over these incredible architectural drawings and models has changed my appreciation for the work that goes into planning parks and playgrounds.
- Link to a sample of an architectural drawing:
https://archive.noguchi.org/Detail/archival/79506
- Link to the park model
https://archive.noguchi.org/Detail/archival/31945
Process
Once I located the datasets, I began by uploading the four files into Carto and created a new map project. Given the scope of my research question and the type of datasets I found, my “approach would be analyzing shapes based on spatial relationships.” In this new project, I first imported the School Location shapefile and the School District shapefile. I created an Intersect and Aggregate Analysis that used the School Districts as the base and the School locations as the target layer based on Count. This process would enable me to create a choropleth map colored by the total count of schools in each district.
Then I imported two more layers: NYC Park & Directory of Playgrounds. I layered these two above the school district in order to visualize and compare the location of these two datasets over the school districts.
Visualization
After the datasets were layered in Carto, I used Vis Palette to inform my selection of colors that would still maintain contrast/distinction for color-vision deficiency. The School District layer was stylized based on the value in 5 buckets creating. Once the base colors (reddish scale) were selected I then used Vis Palette to compare the green for NYC Parks and the Orange points for Playground locations.
A legend was added to represent each layer and a click activated pop up for the School District layer that shows the School District number and the total count of schools within it. Lastly I also changed the based map layer to gray but decided to leave the geographic labels to orient future users.
Reflection
It was such a delight to finally work with visualizing the Playground dataset I first used almost a year ago. Being able to create layers of different datasets in Carto enabled me to create both contexts and attempt to compare aspects such as the distribution of playgrounds in and outside of NYC Parks. I think this map was more successful in comparing playground to park distribution in NYC than visualizing a direct correlation between playgrounds and the number of schools in a district. I did create a second map that colored School Districts by playground count, but the map layer would only be meaningful when compared alongside school location count by district but I couldn’t combine them in less than 4 layers (maximum layers for the free Carto version).
Future directions for this experiment could be exploring 311 complaints about specific playgrounds or parks or further drilling down into the school dataset to determine the type of school (high school middle, lower, public-private) and distances to parks or playground types within a district.