Data Study on Noises in NYC
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Intro
Since when I arrived in New York, what continues to shock me is the noise, especially car noise. Two things that mostly bothered me in Manhattan is the Sirens and the loud car alarms. Now I'm curious if other people feel like I do, so I performed a data study.
311 Noise Complaint
I could find noise complaints reported to the 311 hotlines since 2010 in NYC open data.
But before I begin to analyze 311 data, I wondered how high is complaint level compared to other big cities. Since Seoul has similar complaint report system and the population of two cities are very similar, I've compared total number of complaints filed last year.
Result shows that new yorkers complaints 10 times more about noise than the citizen of Seoul. This shows the severeness of noise pollution in NYC indirectly. And my assumption is that main cause of noise pollution is car noise and most noisiest area is Manhattan.
Data Analysis
To verify my assumption, I checked number of noise incidents against borough plot.
Manhattan was in 1st place without doubt, But Brooklyn is in second place with slight difference and queens in third place. I was surprised to see queens in 3rd place because I did not witnessed any car noise in there.
To see what's happening more clearly, I tried plotting incident coordinates on heat map.
I could see that noise complaints are widespread over NYC, not concentrated in Manhattan.
Since there was "incident description" in 311 noise complaint data, I put it in tree map plot.
The most outstanding category was "Loud music/party". Not "car noise" My assumption was wrong.
I could verify this in the incident against category bar plot too.
And finally incident count against day of week plot revealed the main cause of noise complaint. What people of NYC just could not stand was party noise on weekend.
The app - party noise tracker data
To tackle the party noise problem, or to gain profit from the phenomenon, one can need to find where are the hottest party noise site. So I built the party noise app with which you can pinpoint the recent party noise incidents.
To be more useful, I need to implement daily batch job to make data updated automatically but that is left for future improvements just for now.
About Author
