Show HN: 1,053,713 buildings in New York City, shaded by year of construction | Hacker News Reader

Show HN: 1,053,713 buildings in New York City, shaded by year of construction

bdon.org

96 pointsbdon34 comments

Fraunces Tavern[1], the oldest surviving building in Manhattan, built in 1671 should have been glowing red in the Financial District. In this visualization it is shaded blue for 1900. It was the first building I checked and it was spectacularly wrong; this does not bode well.

[1] http://en.wikipedia.org/wiki/Fraunces_Tavern

Yeah I immediately looked down there also.

I know in the PLUTO data a lot of the entries have a year build and a number of years renovated. Even then though it appears that the year built is being used throughout the map. Probably requires an inspection of the actual row in the PLUTO data containing Fraunces Tavern.

That's concerning, the only two buildings I see in PLUTO built before 1700 are in Queens and Staten Island.

This is a very common issue in NYC, my building is also listed as built in 1900. Someone with real estate knowledge told me a fire destroyed many building records (not sure when), so the city defaulted those records to 1900.

edit: I just read this somewhere: "But a word of caution. Buildings built before 1900 are often labeled 1899 or 1900 in city databases." So there may be two separate issues at play here.

Does the nyc opendata initiative have any crowd sourced programs to repair broken records? Could be a win/win for developers and the city if they would accept feedback on their datasets.

Unfortunately gov has for the most part been reluctant to adopt such practices so far. However NPS Maps have been using crowdsourced data from OpenStreetMap and park visitors in their map products.

That building was actually built in 1719. Which is probably why they picked arbitrary dates vs using 3rd party input. If they see 1900 then the construction date is old and probably unknown which is good enough.

DeLancey built the current building as a house in 1719. The small yellow bricks used in its construction were imported from the Dutch Republic and the sizable mansion ranked highly in the province for its quality.[7] His heirs sold the building in 1762 to Samuel Fraunces who converted the home into the popular tavern, first named the Queen's Head.

Yes, the Flushing Quaker Meeting House is listed as 1900 as well, although actually it was built in 1694.

http://en.wikipedia.org/wiki/Old_Quaker_Meeting_House_%28Flu...

Update as to why the dates are wrong for old buildings: (hint, Y2K) https://twitter.com/technickle/status/378326979164241920

Thanks for finding that out! I wouldn't have guessed.

Why is the entirety of Columbia University labeled as 1901? Low Library was built in 1895 [1], and a number of the buildings on campus date from when the area was a mental asylum [2].

[1] http://en.wikipedia.org/wiki/Low_Memorial_Library [2] http://en.wikipedia.org/wiki/Bloomingdale_Insane_Asylum

It seems to be treating all of the Morningside campus as one building with the address 530 W 120th St, built in 1901. But that address corresponds to CEPSR, which was built in the 1990s. :/

Nice. Are you doing this the same way they did The Netherlands?

I know I threw out the suggestions when I saw the Netherlands map that this could easily be done with PLUTO. Thanks for the great work. I can see my Apartment building was built in 1910.

This is done with two different datasets, PLUTO and the building footprint dataset on https://nycopendata.socrata.com/.

Isn't building footprint included in the PLUTO corpus? Does nycopendata provide better building outlines?

You're right, there's something else called MapPLUTO: http://www.nyc.gov/html/dcp/html/bytes/applbyte.shtml that seems to include building footprints. My bet is they're from the same source.

Right. That's what I was using when I was playing with the pluto data in openGl. Unfortunately/Fortunately some of you other programmers are much better than me at displaying this stuff in web apps.

MapPLUTO's footprints aren't for the buildings, but rather the boundaries of the property. It's a big shapefile for each borough. Totally worth exploring. It's got an insanely detailed metadata breakdown that explains a ton of the eccentricities.

As for the dates, most are guesses. If you plot the dates, you see they're overwhelmingly years that end in 0.

Thanks for the link to nycopendata. There is a plethora of interesting NYC geospatial data there that I was unaware of.

Who wants to mash up this information with the 3d information on buildings in NYC to see how the skyline in NYC has changed over the last century?

Might not work. It's not like the lots were vacant before they put up the most recent buildings.

I saw a mention of something like this at BMVC this week. Check out http://www.cc.gatech.edu/~phlosoft/

There are a lot of interesting outliers:

1834 - 56 Irving Place | 1800 - 30 East 13th Street | 1800 - 140 West 13th Street | 1830 - 28 Jane Street |

Just to name some. Inspecting them on google street view they don't appear to stand out from the buildings next to them.

Also, 65 Jumel Terrace - 1765. Oldest one I've seen so far.

That would be the Morris-Jumel Mansion, an interesting place to visit if you like Revolutionary Era history/design.

http://www.morrisjumel.org/

The museum of natural history was not built post-1995. It's probably giving that as the date because they added the new planetarium in 2000. Which raises the question, is this true for other buildings -- is it showing the date of the most recent renovation of the building (as opposed to its construction date)?

I believe the way PLUTO is organized is with 1) a value for when the building was first built, and 2) up to two most recent dates of building modification. The building modification data is itself worth exploring in some way, if anyone wants to give that a shot.

This map shows the 'first built' column, so it is strange that they would get the museum of natural history wrong.

See http://www.nyc.gov/html/dcp/pdf/bytes/pluto_datadictionary.p...

This is beautiful. Best visualization on top of PLUTO data I've seen so far.

Hrm, I just see a black box: http://i.imgur.com/ZDJzT62.png (Chrome 30, OS X).

Yeah, my server is a bit overloaded. Working on that now.

Reminder to self, triple-check your CDN is actually caching the things that matter!

The Museum of Natural History was built in the early 20th centuryish not 1995. Does not portend well.

Would be even cooler with integrated Google Street View!

NYC is a beautiful place.

Reply on news.ycombinator.com