A tale of two cities

We’ve spent the past few months combing through NYC’s open data in order to get an accurate picture of the heating crisis in New York. Our investigation uncovered a lot: heating complaints have increased in each of the past four years in New York City, and last winter inadequate heat was the number one complaint in four out of the five boroughs.

Lots of people have told us that the only reason complaint counts are up is because the past few winters have been unusually cold. And that’s true; weather obviously plays a role. The colder it is outside, the more heating complaints come in to 311. We saw this when we compared the city’s daily complaint count to historic temperature data from Weather Underground. But even if complaint counts are only up because the last few years have been unusually cold, that still means more people are freezing in their homes. We can’t control the weather, but we can make sure that when it’s cold, more landlords are following the law.

Looking at the data made us wonder what other systemic issues are contributing to such high heating complaint numbers. And so, this week we’re focusing on income.

As you may have guessed, our Coldmap suggests that individuals living in lower income zip codes submit a higher number of complaints. But we wanted to know whether there’s truly a correlation between low income zip codes and higher heat complaints.

Using the U.S. Census Bureau’s CitySDK toolset, we compared heat complaint count and median income in Manhattan and the Bronx. In the scatter plot below, Bronx zip codes are represented in orange and Manhattan zip codes in blue. Our x axis shows the median income for the given zip code and our y axis shows total complaint count.


It’s clear that on average, zip codes with lower median incomes have higher complaint counts. Zip code 10458, located in the Bronx just west of the Bronx zoo, had a whopping 7,726 complaints last year. By comparison, zip code 10007, which encompasses the World Trade Center and City Hall in lower Manhattan, had only 15.

For reference, the median income in New York City is $50,711. (For those who need a refresher, median refers to the middle point in a series of ordered data, while mean refers to the average. Medians are unaffected by outliers at the top and bottom, while means can more easily become skewed).

Of course, we have to employ statistics to demonstrate a true correlation. In this case, we began to look at curve fits. A curve line fit helps us ‘predict’ the location of additional data if we were to plot it, and helps us determine whether a true correlation exists.

In the chart below you can see how the curve fit demonstrates this correlation: our line shows that complaint counts rise with lower median income levels and dips lower as the median income for a zip code increases. If a new zip code from the five boroughs with a lower median income were added to the chart, we can be fairly certain that the complaint count would be higher than a zip code with a higher median income. While our Coldmap and other analyses led us to believe that income levels played a role in heat complaint counts, our analysis this week makes it clear that there is a strong statistical correlation between income level and complaint count. 


I would imagine most of you reading this are saying, “Duh!” right about now. But it’s worth a reminder that adequate heat is protected by law in NYC, regardless of how much money you make or how much you pay in rent. This drastic disparity between who suffers from lack of heat and who doesn’t should not exist. At Heat Seek, we’re doing everything we can to shine a light on this issue. We hope you’ll support us in ensuring that all New Yorkers have access to the decent housing they deserve.