The Energy Use of New York City’s Buildings, Visualized

Researchers have created a detailed map of energy consumption by buildings. But how useful is it?

Mike Orcuttarchive page

February 13, 2012

If policymakers want to reduce energy consumption, first they must grasp what people’s current habits are. That’s the rationale behind a recently published study by researchers from Columbia University. They used publicly available data to develop a detailed picture of how energy consumption in New York City buildings—which account for two-thirds of the energy the city uses every year—varies by location.

Such spatially distributed information, the researchers argue, can illuminate good candidates for cost-effective retrofitting to increase energy efficiency, and can inform strategies for future distributed generation systems. For example, small producers of reusable waste heat or excess solar power could save on transmission by locating buildings near the places that are the biggest users of energy. As more and more detailed information is made publicly available by utilities, appliance manufacturers, and governments, models built from such data will also become more useful. “Information is the first step toward sustainability,” study author Vijay Modi said in a statement.

The group built a statistical model based on several large data sets. Data the city government gathered from utilities provided zip-code level numbers on electricity, natural gas, fuel oil, and steam consumption in 2009. Information on the building floor areas of each of the city’s tax lots came from the city government as well.

More data, from U.S. Energy information Administration, reflects how those energy sources are generally used—whether for space heating and cooling, water heating, and general electricity uses like lighting. The model provides metrics, such as the yearly kilowatt-hours of energy used per square meter of building area in each tax lot, which can be used to understand the spatial distribution of building energy use in New York City. It’s not quite building-by-building, but close—about a million buildings take up 859,134 tax lots.

Below is an interactive visualization, posted online recently by the Columbia researchers, that shows the total kilowatt-hours used per square meter of the land area of each tax lot.

But how useful is this particular map? Comments on the original post raise the valid point that perhaps a map showing kilowatt-hours per square meter of building area would be more useful, since the metric of energy used per square meter of the land area misses variations in efficiency among the city’s tall buildings. One person called the map “basically just a height map of the city.”

Lead author Bianca Howard replied: “We realize that visualizing the map based on building floor area instead of lot area…may be more useful to users. Our purpose for normalizing the energy consumption by lot area was to ensure that large tax lots with few buildings didn’t stand out.”

Keep Reading

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.