Mapping New Orleans

*** This page is a work in progress for a current project, so there are parts that are not complete (yet). Keep checking back for new updates as the project grows! ***

My largest DH project to date is a mapping project that I intend to serve as a companion for my dissertation.

I got this idea a few years ago now, and it became the focus of my coursework during the Graduate Certificate in Digital Humanities that I completed at ASU. When it came to producing something as a part of my culminating project, I worked with Professor Mike Tueller at ASU to see how we could make this map come to life.

In one of our first meetings, we talked a lot about what I wanted this map to be and some of the philosophical questions about mapping.

The task was to create a process for processing and mapping data that:

• allowed me to collect my research data in a single place
• allowed someone reading my dissertation to view relevant details and events visually
• allowed me to create unique visualizations to include in my dissertation when such are not available or to support new research claims
• was easy to add to and regenerate as needed, even every day if the research was proving fruitful
• separated the code of the map from the data of the map as much as possible. The goal here was to create something that could easily be adapted by others, or even myself for reuse.

I had tried, for months, with no success to teach myself ArcGIS, so when Dr. Tueller suggested we create a code using Python and Folium, I was ready to try my hand with it, with his guidance. It ended up being the best tool for the job because of the way it allowed us to separate the data and the code “what gets done with that data to make it a map” quite effectively. Essentially there are four files in this project:

• The main database lists each feature, its descriptions such as dates, text, images, and metadata such as what type of feature it is and which chapter(s) of my dissertation it applies to.
• The features sheet provides the color, icon, and type of geometry for each of the feature classes I am using such as canals, floods, land plots, etc.
• The shapes themselves. These are the geometric units that put my features on the map in the right place.
• The code, written with Folium and Python, that combines these three sheets and builds the markers, features, and view controls onto a map that is saved as a single .html file that can be shared and run without the need for running the code each time.

Here are some of the questions that we thought about during early meetings to consider during the mapping process:

• How do maps impose (false) order?
• How do maps represent or impose chaos?
• Where do we see order and chaos in New Orleans via mapping?
• What limitations or privileging are we imposing by even doing this research via mapping?
• How is common space or public space created and struggled over?
• What new visualizations can we create from qualitative and quantitative data and then combining that with cartographic data?

A Constantly Evolving Process

Originally the map plan would be based around census tracts. For each feature to be graphed that was a polygon in shape (not a line or a point), a spreadsheet of all census tracts of five parishes in southern Louisiana including New Orleans was consulted and marked with a TRUE. The map would load the census tract layers and then apply the features based on the presence of the TRUE maker. Not only did that almost guarantee that some shapes would be in fact the wrong shapes, but this process was tedious and left significant opportunity for human error as there were more than 22,000 census blocks in the sheet and each would be entered individually and manually.

This also made it difficult to keep all of the geographic data for all features (polygons, lines, and points) in a single file because not only did the census blocks not allow for points or lines, but did not mix well with the few .geojson files for some features that were already available online, such as the Laffite Greenway that was once the Carondelet Canal.

So it because paramount, for accuracy, speed, and simplicity of the code, to find a way to create .geojson entries for each of the features. And that is when I stumbled upon the tool click2shp from UCLA. With this tool, I could draw out the features I need, regardless of whether they are lines, points, or polygons, and export them to a .geojson file (and add them to a single file for the whole project) with their unique ID already included in the strong to match it to the main data file.

Still Much to Figure Out

Something I am still working on overcoming when it comes to my shape creation is the variance of hand-drawn historic maps versus today’s satellite maps that I am drawing these shapes on, and of course, because New Orleans sits on a river, the shoreline is constantly changing. Things do not always match up, but I believe I have found a tool for that, and once deployed I might be able to redraw the shapes quickly to make them more accurate. I will be exploring this more in the new year and will add an update here when I have some thoughts on it.