Brooklyn Builds Up

Phase One Project lives here.


Visualizing Brooklyn’s housing boom


The view outside my window.

The sounds of construction in my neighborhood are constant; walking from my front door to the subway involves crossing streets to avoid cranes, ducking under scaffolding, and constantly being on the look-out for falling debris (actually, I should probably be more aware of this than I currently am). I live in a neighborhood undergoing rapid transformation, and for that reason I thought it’d be illuminating to look at NYC’s database of DOB permit issuance.

It’s taken some time playing around with the data to determine what I might be able to learn from it, and what can be used as a visualization. The data originates in 2013 and is updated daily; it contains information about the work permits issued for construction in the city’s five boroughs.

To make it manageable, I pulled information for my area: Community Board 1 in Brooklyn. I believe it will be interested to look at the permits issued for different types of construction (new buildings, and three levels of building alterations) as a way of visualizing the activity happening around my neighborhood. As so many people have been moving to this area, and as so many businesses have been opening, it may also be interesting to separate this by residential and non-residential construction.

A limitation to this visualization may be that the data doesn’t reach back further: my neighborhood was already undergoing change in 2013, and so having information for earlier dates could help me show how drastic these recent changes have been. An alternative method may be to look at permits for each community board district in Brooklyn, and see where the most construction is happening (this will also help with apartment hunting!) It also requires some cleaning; for example, there are multiple instances in which dates like 2013 are mistakenly entered as 2103.

My limited D3 skills at this point may prevent me from making my visualization as complex as I’d like to: for this assignment, I may stick with simple bar charts. My goal will be to visualize the data clearly and intelligently without getting too fancy.

The messy work of creating our first data visualization

Our first exercise was an important reminder that even when you’re working with simple tools, data visualization is complex business — and that even the simplest of datasets can yield eye-catching and informative visualizations.

When asked to visualize something about the makeup of our class, my group’s first instinct was to get creative in terms of what data we chose to collect. This lead to a frenzied 20 minutes of chasing down our classmates and asking them a) What country they’re from, b) What their favorite cuisine is, and c) What their least favorite cuisine is. Problems arose from the start: some people forgot to list one of the three data points, some turned out to claim more than one country of origin, while others had unexpected interpretations of what constitutes a cuisine (“chicken curry,” “American Chinese food,” etc). These responses made it more difficult for us to create our visualization the way we’d planned. We had to keep reminding one another to calm down, lest we make our work even more complicated than it already was.

At one point, early in the process, a group member suggested that we scrap our idea altogether and do what sounded to me like the simplest thing ever: visualize our class by what school/department each student was in. I immediately rejected the idea, arguing that it was too easy and, therefore, boring. In my mind, a data visualization would only be interesting if it began with an interesting concept–thus food preference and nationalities.

I am now reconsidering this (mis)conception, thanks to the group that took the very concept that I dismissed and made what turned out to be one of my favorite visualizations of the class. Yes, their visualization was simple, but it was clear, easy to read, and made good use of color to help illustrate a point (even if I didn’t completely agree with their categorizations of the programs). And it helped me understand the makeup of the class in a way that I hadn’t just by listening to everyone go around and introduce themselves.

In the end, my group ended up simplifying our data to two categories: country of origin, and favorite cuisine. We noticed some interesting, if not particularly relevant, patterns in our admittedly small dataset. As I continue in this class, however, I will remember the lesson that less can be more — and that there are no boring datasets*, only boring data visualizers.

*Okay, I still suspect there may be datasets out there that aren’t worth anyone’s time.