A Data Study on NYC's 311 Noise Complaints
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Please visit my R Shiny App here to view the all the cool interactive graphics in the presentation. To view the original source code, you can visit my Github page here. If you are interested in learning more about my background, you can visit my LinkedIn page here.
Motivation
We've all been through the worse of NYC's noise pollution. Cars honking, loud construction work, next door neighbors blasting music, you name it. In this data case study, we will examine NYC Open Data's dataset on 311 Service Requests (link here) to better understand Big Apple's noise situation, and perhaps to inform the next wave of misophonic NYC dwellers on where they should live so that they can get a good night sleep.
About NYC Open Data
NYC Open Data was initially made available to the public in 2015. As part of the Open Data For All initiative, NYC Open Data aims not only to benefit New Yorkers as it makes once privately held data available to the public to view, it also creates opportunities for people who work with data, such as data scientists, to make creative uses of the data to help New Yorkers and the government make more informed decisions.
What Are 311 Service Requests?
NYC311 provides 24/7 support to New Yorkers by providing a one-stop shop for a multitude of general needs. Within the proximity of the New York Metropolitan Area, callers can dial 311 to request for a variety of governmental services, such as filing a noise complaint, or to request for directions.
The Data Set
The data set used in this analysis contains 311 Service Requests from 2010 to present time, which totals to over 15M observations and about 10GB in file size. Each observation contains 53 variables describing the service request, such as complaint type, date received, incident address, and resolution description. To reduce computational overhead, only eight columns were used in this analysis.
| Column Used | Definition |
| Unique Key | Unique identifier of a Service Request (SR) |
| Created Date | Date SR was created |
| Closed Date | Date SR was closed |
| Complaint Type | Complaint type |
| Incident Zip | Incident location zip code |
| Borough | Incident location borough |
| Year | Year extracted from Created Date |
| Hour | Hour extracted from Created Date |
Data Noise Complaints Takes Silver in Highest Complaint Count
Among a slew of complaint categories, noise complaints ranks 2nd in highest complaint counts since 2010, with approximately 2.1M complaints in total. Some other high ranking complaint categories include heat/water system (1st), street/traffic condition (3rd), general construction/plumbing (4th), illegal parking (5th), and unsanitary condition (6th). These complaints alone account for about 65% of all complaints filed to date since 2010.
Quiet Time Begins at 6AM of the Day
In this illustration, the noise complaints are grouped by the hour of the day using the 24-hour time format. It is apparent that noise complaints on average peak at around the 23rd hour (i.e. 11PM). On the contrary, the 6th hour of the day (i.e. 6AM) accounts for the the least number of noise complaints received.
Visualizing Data on Noise Complaints By Neighborhood
Combining the power of Leaflet, R Shiny, and Cartographic Boundary Shapefiles, we can create an interactive map to paint noise complaint counts on a gradient scale across zip codes. In the R Shiny App, the user can hover over each zip code to show the neighborhood's name, zip code, as well as the total number of noise complaint count for the chosen hour of the day. Below we compare the noise complaint counts between 6AM and 11PM. It is apparent that 11PM experiences a much higher number of noise complaints compared with 6AM.
What Could be Driving All These Noise Complaints?
In the illustration below, we break down each of the top 10 complaint types by their complaint counts annually. At first sight, one can see that illegal parking complaints (i.e. purple line with stars) appear to trend similarly with that of noise complaints (i.e. light blue with pentagons). Both of these complaint types increase at about the same rate from 2010 to 2016. Note that all complaint types' counts 'dropped' in 2017 because we only have data for the first 4 months of the year at the time this analysis was done.
Are Illegal Parking Causing Noise Complaints?
To look deeper into the relationship between noise and illegal parking complaints, we put them onto a scatterplot. As we have seen in the previous illustration, both of these complaints appear to be increasing at the same rate year-to-year. Note that 2017 has been excluded as we do not want to compare partial year results.
When visualizing noise and illegal parking complaints onto maps, it can be seen that that they affect different areas of New York City. Noise complaints appear to affect primarily the areas surrounding Lower East Side, Upper West Side, and west Bronx. On the contrary, illegal parking complaints affect primarily areas of Brooklyn and Queens. While the illustrations below is specific to the 2017 data points collected, the same pattern can be seen for other years as well. We cannot conclude that there is a direct relationship between illegal parking complaints and noise complaints.
Guide to Choosing a Neighborhood to Live for Misophones
Here is a quick guide on choosing a neighborhood to live in if you absolutely cannot deal with noisy neighbors.
Future Research
Given a two-week timeframe for completion of this project, the scope of the analysis was limited. If given more time, it would be interesting to conduct the following research:
- Include the detailed complaint description from the original data set to understand the root causes of noise complaints. Such an analysis will require the use of natural language processing to process the large volume of texts.
- Create an R Shiny app that can fetch real-time data from the dataset, allowing users to visualize any complaint types through the interactive map.
About Author
