Do New Yorkers make more noise than before under the threat of COVID-19?
As I write this, the ceiling above me continues to feel a dizzy shock. Someone is jumping upstairs.
Maybe they’re cranky, maybe they’re exercising, but the only thing I could know for sure is that such bothersome noise didn’t happen so frequently last year, when COVID-19 had not yet struck the city.
Yet New York City 311 service requests statistics kindly informs me that I am not the only person in New York suffering from this problem.
It says:
“New Yorkers are reporting a lot more residential noise cases this year than ever before.”
Nevertheless, we cannot yet hold sure whether the increase in the noise Report is directly related to the COVID-19 epidemic.
To judge intuitively, however, to remain a social distance and to encourage self-isolation may increase residents’ length of stay at home, while the rapidly rising unemployment rate along with a huge expenditure on disinfection will place greater pressure on residents’ life. Hence, it sounds fairly reasonable to expect a rise in the frequency of emotion venting — both referring to the making of noises and the reporting of noises.
Despite the continuing high unemployment rate, the 311 residential noise complaints all over the year present an uneven distribution pattern in a week. Fewer complaints tend to occur in the middle of a week, while the noise reported on Sundays almost double that of a weekday.
Among all reported residential noise complaints through NYC 311, most were in regard to loud music and party (264,928 in total). Banging and pounding ranks the second (83,162 in total).
Not all type of residential noise reported shows the same imbalanced distribution within a week as the total statistics. Loud music and party tend to happen and be reported during the weekend, which is evident, while the complaint records of banging and pounding present smaller difference between weekends and weekdays.
Such fluctuating pattern of the NYC’s residential noise complaints contributes to the difficulty for any attempt to test its correlation with other factors, such as the COVID’s daily confirmed cases and death cases. To exclude extraneous variables as much as possible, I referred to two subsets of the NYC 311 data as the experiment group and the control group respectively, which are the residential noise complaints reported from March 2020 to October 2020, and from March 2019 to October 2019. This selected time range is specified by the limitation of COVID’s daily data, which started documenting merely since early March this year.
According to the observation over the past year’s statistics (shown in the first graph), before the year 2020, the number of residential noise complaints appears little change from year to year, growing slowly and steadily. Therefore, to interpret the net-difference before and after the pandemic, I set two data groups from two adjacent years to compare. As a complement, I also evaluated the daily count of null requests (when there were actually no need to call the police), daily no-entry ends(the police can’t enter the reporter’s building), daily party noise reported, and daily pounding noise reported in both datasets. After doing this, I sent everything to a test for correlation.
Sad though, the result from correlation test shows scantly any evidence to confirm the relationship between COVID crisis and the residential noise complaints, which was unbelievable.
However, when turning to the help of visualization, we could finally figure out the reason behind the test failure: there was a time lag between the potential cause and effect - that is, between the pandemic and the noise report increment.
For one thing, we could see the huge difference between this year and last year’s complaint data: they differ not only in number, but also in shape. For another, the evident rising peak of this year’s residential noise complaint data simulates a bell curve, which seems to echo the pandemic curve of confirmed cases, although both curve went shaky and fluctuate drastically.
Beyond the main conclusion, for which the analytical steps would still need to refine, a few sarcastic findings, though, are worthy of noticing. When checking the peak points of the reported residential noises, I realized that the increment from last year were mostly (for some even totally) contributed by null requests. Guess this could be due to hyper-vigilance or nervousness, or probably triggered massively by particular event happened on that day. Another surprising observation from the result was the scarcity of no-entry occasions for police officers (compared to other types of complaints’ responsive personnel). The no-entry statistics were so stable and almost hold same with the past year’s, only presenting some fluctuations during the pandemic reported peak.
Thank you for reading.