Total Annual Building Energy Consumption for New York City

February 5, 2012 8:30 PM

An amazing piece of statistical analysis produced this zoomable (down to the block level) map of energy consumption for New York City, based on Spatial distribution of urban building energy consumption by end use. [via]
posted by unliteral (30 comments total)

Any NYC Mefites want to confirm the accuracy?
posted by unliteral at 8:30 PM on February 5

I must have left the light on.
posted by jonmc at 8:42 PM on February 5 [1 favorite]

Any NYC Mefites want to confirm the accuracy?

Well, it doesn't look entirely unlike a map of how tall the buildings are, which is unsurprising, when you think about it.
posted by planet at 8:43 PM on February 5 [6 favorites]

Shouldn't the energy companies be able to produce the actual map after a very simple geo-coding exercise?
posted by one_bean at 8:47 PM on February 5 [2 favorites]

Privacy in energy meter data seems to be a real hot-button topic. I can see why the academics tried to get started without actual data.

In a world of warrantless wiretapping and pervasive internet surveillance, it seems kind of quaint and naive that people get so upset about their home's electricity/gas consumption being disclosed by the utility companies. But so it is.
posted by anthill at 8:52 PM on February 5

Is this anything more than a population density map at this point? They need a way to correct for people-per-building.
posted by Defenestrator at 8:53 PM on February 5

Well, I'm not able to get in and see the full text, but it seems from the description that they are using very high-level data, like at the zip code level, as their source for energy consumption.

Then, they are taking a block and tax lot map and coloring in the lots and blocks based on their estimated usage from that high-level data.

So, if zip code 10001 uses 400kwh of energy, and there is a big building that is 1/100th of that size, then it is using 1/400kwh.

That is what my quick post-superbowl and a few beers deep reading gives me.

The problem with that would be if there was a big building leaking energy like a sieve (or if jonmc left the lights on at Yankee stadium), you wouldn't see that loss as it would be distributed over the whole zip code. It doesn't make the project unimportant, it would just make it less useful at such a zoomed in level. Which is what everyone is going to do from NYC -- you are going to zoom in on your building.

Also, I am curious about their usage breakdowns because it seems like the cooling costs are very low. For example, I have my AC on right now doing battle with my evil steam radiator. Since, I then have to leave my AC on year round, I would think my cooling costs would be higher. So, I wonder when their survey was taken. Was it an annual estimate, or was it taken in the middle of the winter. And clearly not this winter.
posted by This_Will_Be_Good at 8:53 PM on February 5 [1 favorite]

Their statistical model utilizes zipcode-level energy consumption data to estimate the average annual energy use for every tax lot—at practically building level—through all five boroughs of the city.

It be nice to know what this model is, but I can't access the paper. My instinct is the same as This_Will_Be_Good's, in which case this isn't difficult at all.
posted by desjardins at 8:56 PM on February 5

Oops, not population density, but building size I guess. So what it needs is a way to easily compare against buildings of the same size (which isn't possible yet based on the way they obtained the data).
posted by Defenestrator at 8:57 PM on February 5

What they need to do is take those numbers and divide by however many people live and/or work on each block.

Then they'll have something worth looking at.
posted by Sys Rq at 9:11 PM on February 5

What they need to do is take those numbers and divide by however many people live and/or work on each block.

I don't think they necessarily need to do that. If you want to look at the efficiency of buildings in terms of energy use, it isn't always a population question. For example, I would be curious to see what the energy usage is for One Times Square. That is an example of a building where NOBODY lives or works, but is just blazing with electricity all the time, though their steam usage has probably gone down now that they took the Cup Noodles sign down.
posted by This_Will_Be_Good at 9:19 PM on February 5

The technology that's described in anthill's article would give the energy company the ability to control your energy usage at times of peak load. I'm all for smart grid, but even I get nervous about that. I mean... Pepco? Trust?

On the other hand they already have data on energy consumption of individual homes. If they asked my permission to put it up on a website where I could see how much I used compared to other people who agreed, I'd probably be OK with that. Especially if the historical information were available to potential home buyers, who could see how much it really cost to keep the house they were thinking about warm. Historical price information is public and online.

Interestingly, one of the main reasons utilities subsidize energy audits is so that they can collect information about energy consumption. Also a trade-off I'm willing to make. But definitely a close call.
posted by RandlePatrickMcMurphy at 9:22 PM on February 5

If you want to look at the efficiency of buildings in terms of energy use, it isn't always a population question.

But how can you compare the efficiency of two different buildings in terms of energy use if all you know is how much energy each building uses?
posted by planet at 9:30 PM on February 5

Interesting that there's no data for 33 Thomas Street (formerly the AT&T Long Lines building).
posted by mullacc at 9:33 PM on February 5

Any NYC Mefites want to confirm the accuracy?

As others have mentioned, this is pretty much a map of how tall the buildings are. This is illustrated pretty well when you look in the Northeast Bronx, where I grew up. Co-Op City is listed in the 200-300 range, which looks pretty bad except when you consider that it consists of 35 high rise residential buildings, with heating, cooling, and hot water to the entire complex all supplied from a central plant.

Then, look a mile to the east at City Island (in the 50-75 range), which is mostly pre WW2 single-family construction, with all the leaks and drafts that go with it. There's just no damn way those houses are more efficient than the ones in Co-Op City - the only explanation for the difference is the density per square foot of land.
posted by deadmessenger at 9:37 PM on February 5

I believe the dark block-long chunk in Chelsea between 8th and 9th Avenues is the building that Google bought.
posted by A dead Quaker at 9:46 PM on February 5

Ok, what's really interesting is that when you look at Co-Op City (my old neighborhood), about a quarter of the complex (Section 5, for those who are familiar with the area) is in a higher energy-consumption category than the rest, which strikes me as more than a little odd. The buildings in that section are identical in every way to the others. They're fed from the same heat/hot water/cooling system as the rest of the complex, the floorplans and density are identical, the same builder, the same materials, the same construction techniques - they're the same buildings. It just makes no sense at all.

To me, this discrepancy calls into question their entire model.
posted by deadmessenger at 9:47 PM on February 5

But how can you compare the efficiency of two different buildings in terms of energy use if all you know is how much energy each building uses?

Well, that is sort of what they are trying to do here -- they are trying to make comparisons by building function.

So, if two similar-sized residential buildings are using vastly different amounts of energy, that may be an indicator that someone may be using a bit more than their fair share of the grid (for example, a secret Manhattan project lab in Pupin Hall), or that someone is really doing things efficiently.

They have the building functions derived from tax lot data, it just isn't super detailed, and also doesn't really seem to account for all the mixed-use buildings (for example, did they pool the energy surveys by building? by person? by owner?).

Having data at the building level can be super helpful. For example, look at the water system. From the source to the end user, they know that something like 50% of the water is lost (I forget the actual number, but it is high). If you can plug all those leaks, it would be great. You could think of this energy map that way if it was super accurate. If you could see building by building within a given function, and a few were WAY off, it might be worth a look.

I pass by an old building a lot where I can see down through the wire glass into a creepy basement...and there is a boiler in there that looks like something from Dr. Frankenstein's lab. It is so large that I assume it was carved out of the Schist it rests on. It has to be 150 ft. long. I am assuming that it is not LEED platinum.
posted by This_Will_Be_Good at 9:47 PM on February 5

it seems from the description that they are using very high-level data, like at the zip code level, as their source for energy consumption.

Try the individual building level. I clicked on the map proper and zoomed in, and found that you can hover the mouse over a spot and get the stats on a specific building.
posted by EmpressCallipygos at 2:42 AM on February 6

Eventually the building code will look up in the sky and see a huge hot thing. And on that day, our energy use will go down by about 50%.
posted by DU at 4:23 AM on February 6

Privacy of electrical usage data is quite probably about large data center security.
posted by sfts2 at 5:47 AM on February 6

Huh, kind of a technology demo showing what sort of information could usefully be available. Unfortunately, although we have some data averages on energy consumption by building type (thanks Energy Star), it really isn't useful as a way to look at things at the building level. That's both because the variance is huge (I've run into consumption between equal sized buildings that varies by factors of more than 10) and because the data is really only sharp for single-common-use buildings such as an elementary school or a stand-alone restaurant. When you get into buildings with diverse uses that statistical numbers just don't tell you much.

But, ubiquitous smart metering would give us excellent data at this level. Even more important, it would allow energy consumption to be a more constant part of the real estate transaction making real efficiency improvements a more certain add to baseline capital value.
posted by meinvt at 5:47 AM on February 6

Hey, also this uses MapBox for hosting and was designed in TileMill, two projects that I (and many others) have been working on. So it's made of open source things! Modest Maps, Wax, TileMill, node.js, etc.
posted by tmcw at 6:04 AM on February 6 [2 favorites]

There's nothing wrong with this kind of analysis, but it's really forest/trees in terms of energy consumption.

There's a decent point (above) in response to calls for population density consideratinos. While it's true that energy efficiency gains are there in terms of building performance and energy metering technology, this is just loose change in the couch in terms of energy savings.

This whole piece reminds me of a lot of the projects I see coming out of MIT around more 'efficient cities'. The elephant in the room is that our cities are relatively efficient already by their dense design. Either in terms of efficiency per person or per dollar, cities like NYC kick the crap out of most of the rest of the country. This isn't just because it's New York, either. All of the similarly dense cities are orders of magnitude more productive than their alternatives.

If (the US, for example) wants to be more energy efficient, it needs national level controls to ensure that future growth goes into designing places like NYC and not like Phoenix.

Oh wait, the house is looking at cutting federal transit funding...
posted by Reasonably Everything Happens at 6:55 AM on February 6 [1 favorite]

Is this anything more than a population density map at this point? They need a way to correct for people-per-building.

This is interesting as it shows an energy consumption per unit of built area (including the height). Normalizing by population density is relevant if you want to show how buildings with a similar function behave, but I think if you want to show overall energy consumption in a city this is a good measure.

If anyone wants access to the paper, if you memail I will send it on. (I hope it's ok to post this offer here. grumble grumble elsevier preventing the sharing of knowledge)
posted by a womble is an active kind of sloth at 6:58 AM on February 6

Any NYC Mefites want to confirm the accuracy?

I obviously can't confirm the accuracy of this, but I can give you my takeaways from it:

The more "yuppified" neighborhoods and places with restaurants seem to draw more power than average in Brooklyn. For instance, Park Slope (once you get away from the big buildings near the park) is largely single-family brownstones, or often single family brownstone + another family in the Garden Apartment. Many are just as red as places with big apartment buildings with multiple units on every floor. Same goes for the other "upscale" neighborhoods in Brooklyn.

But this isn't too surprising. It would seem that, apart from restaurants, building size goes along with power intake without regard to the number of people living inside of them.
posted by Threeway Handshake at 9:24 AM on February 6

Based on the big red building across from my own building on the map -- which i know is a school -- I'd wager that the explanation for why "building size = power intake" is: the more floors you have in the building, the more hallways you have to have lights burning in 24/7.
posted by EmpressCallipygos at 12:04 PM on February 6

the more hallways you have to have lights burning in 24/7

That, or air conditioning.
posted by Threeway Handshake at 2:21 PM on February 6

That, or air conditioning.

More like and air conditioning. All that heat from incandescent bulbs has to get taken care of somehow.

Then in the hallways, non-adjustable steam radiators forcing tenants like This_Will_Be_Good to prop the window open (I'd never heard of running AC before).

The fruit is hanging so low it's hitting us in the teeth.
posted by anthill at 4:23 PM on February 6

All that heat from incandescent bulbs has to get taken care of somehow.

No, I mean that those kinds buildings will have AC at all. Many, many people in NYC go without it. Only new places have central air, and window ACs are not just expensive to buy up front, but very expensive to run.
posted by Threeway Handshake at 8:11 AM on February 8