One of the best leading economic indicators with regard to real estate markets is construction permits. Construction permits can tell you where new projects are being built, how big those developments will be, what type of occupants those projects are for (commercial, residential, etc.), and a whole host of other insights.
Using this information, real estate developers and investors can project changes in supply. By understanding changes in supply, inquiring minds can get a better sense for how rents are changing and how neighborhoods are changing.
These are the biggest new building permits approved in New York City this August 2016:
This is built using the New York City Department of Buildings permit data which gets released on a monthly and weekly basis in Excel format. You can find this data here. What’s particularly great about this source of information is that it is timely. The PLUTO data that I love so much is only released yearly. NYC DOB permit data is released weekly.
In this post, I will explain the methodology I used to analyze permit data to visualize where the largest new developments currently in the pipeline are located.
The specific dataset I am using is located here (this will directly download the August 2016 zip file hosted on nyc.gov).
Due to computing constraints, I am limiting the permit data to the following rough locational boundaries:
This is important to note because the NYC DOB permit data is not geocoded, so I have to manually geocode it based on what we get from the DOB.
Data Analysis Sidebar: How to geocode non-geocoded New York City data
The best way I have found to manually geocode NYC data is to merge the non-geocoded data (permits) with geocoded data from PLUTO on a common field that accurately parallels location. Luckily, New York City’s borough-block-lot (BBL) system works perfectly.
Each line in PLUTO has a unique geocode (“the_geom”) corresponding to a unique borough-block-lot code (“borocode,” “block,” “lot”). And almost every other city datasource uses some form of BBL as an identifier. Depending on the source of data, use concatenations and your preferred lookup technique (index/match, vlookup, SELECT/WHERE, etc.) to pull geocodes from PLUTO into whatever datasource you’re looking at by matching BBLs.
In this case, I merged geocodes from PLUTO into the NYC DOB permit data, but the technique applies to pretty much every source of data New York City publishes. As you’ll see shortly, this allows me to map NYC DOB permit data through cartography software such as Carto.
The difficulty here is that the PLUTO data is enormous if you want to look at the entire city. There are millions of lots which translates into gigabytes of data even if you’re just downloading boro-block-lot information and no other tax lot data. That’s why most of my analyses are focused on Manhattan and a little bit of the outer boroughs. My Macbook Pro simply doesn’t have the processing power to handle all that data. (If somebody wants to donate a supercomputer or cloud-based solution to the data analytics cause then hit me up on LinkedIn.)
Back to the permits: Where are the biggest new projects?
Based on the data, these are the biggest developments that were approved in August 2016:
You’ll notice that the tallest new building (NB) permits are all in Manhattan, but there are a few projects in the Bronx that are very large from a square-footage basis.
The biggest project by far is 515 West 42 Street in Manhattan. A quick Google search confirms that this is a huge 350-unit residential building. Notice that The Real Deal posted this article on August 16 and they refer to the Department of Buildings data. This supports my earlier notion of the timeliness value provided by this weekly datasource.
You’ll also see a number of permits that don’t have any height or size information. This has to do with permit types. It appears that general construction permits will have this information, but structural and foundation permits will not. At this stage, I admittedly do not know a ton about the stages within the permitting process, but it’s something worth examining in greater detail in the future.
Now, here’s what we can create using the geocoding technique above:
Each dot represents a new building permit and the size of the dot represents the magnitude of the square-footage as stated on the permit. I factored out permits that do not have square-footage information and permits that were disapproved.
Again, please be aware that I cut the data off around the “e” in the “The Bronx” on that map. There were surely permits issued east of the “e,” but I cannot process all of the data with my current setup, so I focused more closely on Manhattan.
The value of this type of analysis is twofold. First, you have a visual aid that accurately represents where new developments are happening. Second, you have official quantification of both supply and type.
Building out this analysis over time and forecasting construction lengths (or finding estimates in public submittal data, if it exists) would give developers significant insight into the economic direction of the New York city real estate market. Add that to the list of things I need to do!