XXIX.3 May - June 2022
Page: 32
Digital Citation

Sounds of New York City

Graham Dove, Charlie Mydlarz, Juan Bello, Oded Nov

back to top 

More than 70 million people in major U.S. cities are estimated to be exposed to noise levels considered harmful by the Environmental Protection Agency [1], and noise pollution is consistently among the most-cited and highest-impact quality-of-life issues for residents. For the past five years, we have been conducting research for Sounds of New York City (SONYC) [2], a large-scale noise-monitoring and mitigation project in New York City, the largest and most densely populated city in the country. We have developed and deployed a network of noise-monitoring sensors, connected with residents affected by problem noise, and built relationships within municipal agencies and local government, such as the New York City Department of Environmental Protection (DEP), which along with the New York City Police Department (NYPD) is tasked with responding to noise issues. In NYC, noise mitigation is based on a municipal noise code, developed and refined over several decades. The code limits noise based on measurements taken at a fixed distance from the noise source compared to either an absolute value or an ambient noise level. Limits vary by activity, for example, construction, music from bars, street noise, and so on. The city's noise code also requires that measures be put in place to mitigate noise from these activities, such as baffle boards on construction sites and sound limiters in bars. This reflects a desire to mitigate noise at its source, rather than placing the burden on those affected. However, because noise is not systematically monitored, mitigation and enforcement is in practice driven by residents' complaints to the city's 311 system for reporting nonemergency issues.

back to top  Insights

Noise is typically transitory, but its impact is often ongoing. Focused sensor deployments can help bridge the gap between limited municipal resources and residents' lived experiences.
Scalable sensor deployments should be flexible and responsive to identifiable issues.
Engaging with the process of mitigating environmental concerns, such as problem noise, requires building relationships with civic authorities as well as residents' groups.

back to top  Background

One key focus of our research has been noise from construction, particularly construction that takes place at night or on weekends and therefore outside of normally permitted times. In NYC, 75 percent of the construction noise complaints made to the city's 311 system between 2014 and 2016 involved work being carried out under permits for construction outside of normal hours [3]; the number of these permits increased from 31,569 in 2012 to nearly 67,000 in 2018. Among these sites is a core of high-impact, repeat cases where construction noise at night and on weekends becomes a chronic problem for nearby residents. For example, the Hudson Yards real estate development is among the largest and most expensive private construction projects in U.S. history. Work on this project was initially permitted seven days a week, and would at times continue almost 24 hours a day, disturbing the communities residing in adjacent neighborhoods. Yet while the circumstances that have caused this problem noise have been ongoing, the sound events themselves are often transitory, making it extremely difficult for NYC DEP officers to be on-site and take timely measurements. Indeed, inspection visits may not be scheduled until many days after the report to NYC 311 was made. Because of this lag between noise and inspection, gaps in enforcement can become a major source of frustration.


In June 2019, we were contacted by the NYC Mayor's Community Affairs Unit about whether we could help residents monitor noise from the Hudson Yards construction project. Later that year, we were also approached by community health researchers from the NYU Center for the Study of Asian American Health who were concerned about how noise from construction and traffic affects the older Asian-American residents of a housing development in Manhattan's Chinatown. This residence of eighty-eight affordable units is home to extremely low-income senior citizens, most of whom are between ages 85 and 100. It also shares a wall with the Manhattan Detention Complex, which is scheduled to be demolished and then rebuilt to hold more than 1,400 inmates in a 300-foot-tall block. This project, which is part of the program to close the notorious prison on Rikers Island, will take at least 10 years.

Beyond construction noise, we have also been approached by residents' groups concerned with noise from the streets outside their homes. Between 2010 and 2015 there were 236,953 calls to NYC 311 about street noise, which resulted in 1,384 summonses and 237 arrests [3]. In September 2020 we were approached by residents of Washington Heights and Inwood, two adjacent neighborhoods in upper Manhattan, who had been experiencing a major increase in noise from fireworks, loud music from parked vehicles, large groups of dirt bikes and all-terrain vehicles, and drag racing in cars with mufflers modified to sound like gunshots. It has been suggested that reports about street noise may reflect other local tensions, are more frequent at the boundaries between otherwise homogeneous communities [4], and that the culture clash of gentrification is a contributing factor. However, it was our experience that the people reporting this noise were diverse, and as likely to be long-term community members as newer residents. For these communities, restrictions imposed as a response to the Covid-19 pandemic may have been an exacerbating factor, as complaints to 311 increased significantly during the summer of 2020.

back to top  Problems With the Current Response to Noise

Our research has highlighted how noise is often just one facet of complicated issues that pose wicked problems. For example, residents around the Hudson Yards development see benefits coming from the investment, and have consistently attempted to offer compromises that would enable construction to take place well outside of the typical workday. Yet 24-7 construction has had a significant impact on quality of life; and the feeling remains that priority is given to commercial concerns and traffic flow. As one resident explained, "I was realizing that at 11 o'clock at night on Sunday, the lights would turn on and you'd start hearing all these construction workers reporting to work. And it was kind of unbelievable. It definitely changed my sleeping pattern, because I had a really hard time sleeping." These situations highlight a challenge for municipal noise codes. Each individual enforcement measurement may be within allowable limits, but complaints reflect the lived experience of people facing long-term work taking place seven days a week at all hours. It seems reasonable that this be taken into consideration.

Another example that highlights the tension between how noise is experienced and how noise code enforcement requires it to be measured and quantified can be seen in construction work that takes place in the street to repair or replace utilities. Noise from this work is allowed to reach 85 decibels whether the work takes place during the day or at night. From an enforcement perspective, levels below this figure may be an annoyance but are typically not a violation. However, as one resident explained to us, this can be frustrating, leaving them asking themselves, What are you talking about? My whole place is rumbling. We found that, in part because of experiences like this one, residents often attempt to quantify their own experience, for example, by using handheld domestic noise meters or smartphone apps that can help them better understand how the noise they experience relates to the city's noise code. Data from these devices, however, typically lacks value in the eyes of municipal officers because it cannot play a role in enforcing noise code violations.

We have also seen how residents self-organize into groups in response to problem noise. They do this for mutual support and to amplify the impact of their noise reports and advocacy. Residents around the Hudson Yards development formed a block association, while residents in Washington Heights and Inwood formed a task force and organized through a Facebook group with more than 1,200 members. Organization also takes place on a smaller scale, such as within individual buildings. This enables residents to coordinate their reports to NYC 311, which brings problem noise to the attention of the authorities and is often viewed as a way of maintaining visibility for ongoing problems. It also supports residents in making coordinated outreach beyond the 311 system to elected officials as a way of trying to escalate their concerns. Anonymized data from 311 reports is available through NYC's open-data portal, which Washington Heights residents used to compare their own record of calls made with those publicly recorded in published data. This helped them understand that the vast majority of reports were simply closed out within 10 minutes, in this instance by the local NYPD precinct.

back to top  Opportunities for HCI Research and Design

HCI research into monitoring environmental concerns has traditionally focused on using sensing technologies to increase individual awareness and understanding or to support community-based advocacy. Our work extends this into direct engagement with municipal mitigation action, requiring relationships with enforcement agencies as well as with residents and community groups. We might think of this as "civic sensing," as it incorporates different aspects of participatory sensing [5] and digital civics [6]. Initially our approach was to deploy fixed-position sensors in public locations like municipal and university buildings or to use urban infrastructure such as streetlight poles. While this enabled us to answer engineering questions regarding sensor design and data collection, such opportunistic deployments are not flexible enough to respond to the needs of residents in dispersed locations across the city, or the municipal agencies that respond to chronic noise problems. Our approach is now changing. To gather data that provides a temporally rich picture of the acoustic environment at problem sites, we are piloting a program in which novel domestic sensors we designed are deployed for periods of two to four weeks with residents identified by the NYC DEP via 311 reports of ongoing noise problems. The involvement of the DEP in this process carries the implication that official mitigation action will result where violations of the noise code are detected. Figure 1 shows one of these sensors mounted on a residential window. Sensor data is made available to residents through a Web app and shared with municipal authorities to help plan enforcement and mitigation activities. The Web app (Figure 2) incorporates real-time noise measurements, enables detailed reporting of noise disturbances to notify authorities, and helps residents find out about construction permits and prior 311 reports or share noise reports with neighbors. In addition to monitoring sound-pressure level, we are incorporating human-in-the-loop machine learning to move toward automatically identifying the source of these noise events.

Photographs of a SONYC domestic noise monitoring sensor installed on the outside of a residential window, with insert showing the sensor in closer detail. Figure 1. Domestic noise sensors are deployed for two to four weeks.
Interface designs from the web-based platform we have designed for residents to use in conjunction with sensors, and which supports our model of civic sensing. From left to right, top to bottom: real-time noise monitoring, reporting problem noise, using open data to identify prior reports and construction permits, sharing report details with other community members or elected representative. Figure 2. Screenshots from the Web app for domestic noise sensors.

Through this research we have identified several design opportunities for HCI. These include: mediating residents' and officials' assessments of data quality; supporting residents' coordinated action, including action with civic open data; and supporting greater accountability. Assessments of data quality vary between residents and officials. Residents' measurements may be discounted as not useful for enforcement, while enforcement officers' spot measurements can seem out of step with residents' lived experience. When sensors are deployed longitudinally with residents, however, it can help their understanding of how noise experiences might be quantified; when this data is shared with the DEP, it supports the scheduling of enforcement officer visits to most likely coincide with when noise happens. Sharing information about reports of problem noise also supports residents' coordinated action, allowing patterns to be identified and issues of concern escalated to elected representatives or municipal officers. Information from civic open-data sources can further empower residents in two ways: in supporting comparative analysis of how authorities respond to 311 reports and in providing access to information around construction permits and variances to these that allow for nighttime and weekend work, or applications for liquor and entertainment licenses. We also see an opportunity to increase trust and support accountability through providing accurate, attributable, and persistent evidence of activities by and for community members and municipal officers, based on open access to anonymized 311 report data that includes attributable reasons for follow-up actions and resolution.

back to top  References

1. Hammer, M.S., Swinburn, T.K., and Neitzel, R.L. Environmental noise pollution in the United States: Developing an effective public health response. Environmental Health Perspectives 122, 2 (Feb. 2014), 115119.

2. Bello, J.P. et al. SONYC: A system for monitoring, analyzing, and mitigating urban noise pollution. Commun. ACM 62, 2 (2019), 68–77.

3. Office of the New York State Comptroller. Noise in New York City Neighborhoods. 2018; https://www.osc.state.ny.us/sites/default/files/reports/documents/pdf/2018-12/health-noise-in-nyc-2018.pdf

4. Legewie, J. and Schaeffer, M. Contested boundaries: Explaining where ethnoracial diversity provokes neighborhood conflict. Amer. J. Sociology 122, 1 (2016), 125–161.

5. Burke, J.A. et al. Participatory sensing. UCLA: Center for Embedded Network Sensing, 2006.

6. Olivier, P. and Wright, P. Digital civics: Taking a local turn. Interactions 22, 4 (2015), 61–63.

back to top  Authors

Graham Dove is a human-computer interaction design researcher currently working as a research assistant professor at the Center for Urban Science and Progress (CUSP), New York University (NYU). His research bridges design studies, human-computer interaction, creativity studies, and data science to explore methods for understanding human-data relations and collaboratively designing pleasurable and meaningful products and services. gd64@nyu.edu

Charlie Mydlarz is a research assistant professor at CUSP and the Music and Audio Research Lab (MARL), NYU. He works on the development and implementation of cyberphysical systems for large-scale, high-resolution soundscape data capture from urban environments with a focus on acoustic sensor development using digital MEMS microphone solutions. His expertise lies in: acoustics, mobile application development, acoustic ecology, digital signal processing, and quantitative data analysis. cmydlarz@nyu.edu

Juan Pablo Bello is a professor of music technology and computer science and engineering, and director of NYU's Center for Urban Science and Progress (CUSP) and Music and Audio Research Lab (MARL). His expertise is in digital signal processing, machine listening, and music information retrieval, and he leads research on sound and music informatics. jpbello@nyu.edu

Oded Nov is a professor in the Department of Technology Management and Innovation at the Tandon School of Engineering at New York University. His research interests include human-computer interaction, the future of work, digital health, interactive technology and decision making, computer-supported cooperative work, and citizen science. onov@nyu.edu

back to top 

©2022 ACM  1072-5520/22/05  $15.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2022 ACM, Inc.

Post Comment

No Comments Found