Published: Jul 26, 2024 | at 03:30 AM
Hi! I’m back with a supplement to a piece in the New York Times’ Street Wars series, which is all about the battle for navigable urban space in New York City. In this article, I’ll go more in-depth about how I devised & computed the claustrophobic metric. We plan to explore this metric further through an in-progress research paper; keep an eye out for that later this year! Let’s begin.
New York City is a large place; almost 469 square miles of pretty dense civilization. Within the city, there are thousands of miles of sidewalks. As you walk through different neighborhoods, you may experience a variety of different atmospheres. In Cobble Hill, Brooklyn, it’s quaint and quiet. In SoHo these days, there are so many pedestrians that they spill off the narrow sidewalks. While a neighborhood’s atmosphere is, of course, a function of time, it is possible to get an average consensus of how ‘crowded’ each neighborhood feels by averaging over time. When we say ‘crowded’, we mean not just with people; we also mean with static objects, or street furniture, or, to get even more colloquial, ‘clutter’. When we mix ‘crowdedness’ within the narrow environment of NYC’s sidewalks, we endeavor to call this feeling ‘claustrophobia’, a direct mapping to the definition in psychology.
Now, we’ll discuss how the metric of sidewalk ‘claustrophobia’ was calculated. Then, I’ll talk briefly about how this metric might be interesting and useful to a variety of different stakeholders.
Methodology - Segmentization
We start with the official planimetric database of NYC’s sidewalks from NYC OpenData at this link. However, the geometries for each sidewalk here are stored as multi-polygons, instead of at the per-segment level. Further, the geometry can be quite complicated, in fact, overly complex for the purposes of our analysis. To mitigate these problems, we perform the following:
Simplify geometry using Shapely library Here, we first simplify the sidewalk geometry to reduce some of the complexity in the street network. We visually inspect several different neighborhoods and find that this minimally changes the shape of the network while moderately reducing the number of points after segmentization. Segmentize points along sidewalks at least every 50 feet. Then, we segmentize the simplified sidewalk network. Segmentization is a process that evenly samples points along each sidewalk, at a predetermined threshold. We use a threshold of 50 feet to balance computational complexity and storage constraints with accuracy.
Methodology - Bringing in Clutter
For this computational analysis, ‘clutter’ is anything that takes up space on the sidewalk. Narratively, some clutter is aesthetic or unminded by pedestrians (like trees, most seating); then, things like scaffolding are denotatively and connotatively ‘clutter’. To identify different types of clutter, I took walks around several different neighborhoods in Brooklyn, Queens, and Manhattan, writing down the different things that I saw. At this point, I tried to match each type of street furniture I saw with a dataset on NYC OpenData, which is a great, official portal that stores hundred of city-related datasets from dozens of city agencies like the Department of Transportation and NYC Parks. To save space, I list all of the datasets I used in next section’s table, along with an access link.
We assign points to different clutters with spatial joins. For each point, we add a buffer (think of this as a larger ring, centered at the point) of 25 feet. These buffers act as a net, ‘catching’ nearby pieces of clutter. Multiple points may count a piece of clutter as ‘theirs’ if the clutter is within both points’ buffer area.
... continue reading