2020 Capstone Projects

Featured Research

We spoke to several of this year’s capstone teams to understand more about their research, and how their projects could help improve urban life, both in New York City and other cities around the world. Learn more:

The Urban Dynamics Around the New York & Milano Fashion Weeks

Sponsor: Politecnico di Milano

CUSP Mentor: Enrico Bertini

CUSP Students: Khouloud El Alami, John Thomas Hughes, and Di Xin

Politecnico di Milano Students: Francesco Alongi, Antonio Esposito, and Andrea Pronzati

Project Website: https://datalife2020.github.io

As a semi-annual event, Fashion Week succeeds in gathering the interest of a massive audience from all corners of the world. New York City and Milan have long been at the center of this event as well as the fashion industry as whole. This research seeks to determine how this event affects peoples’ behavior in the online world by looking at the 2020 Fashion Week event in each city, focusing on what insights Fashion Brands can draw from this in order to better engage and understand their audience, and ultimately developing a comprehensive view of the events as a whole from the perspective of these brands, in the form of an interactive tool.

We will observe the various communities that arise out of the diversity of interests garnered from the unique ideas that exist in the fashion ecosystem. For our analysis of the different interactions and behaviors of these people and communities that come into play with regards to fashion, we will turn to social media data from Instagram and Twitter and employ several machine learning techniques such as clustering, topic modeling, and sentiment analysis.

Security Analysis of Trajectory Data

Sponsor: NYU CUSP

CUSP Mentor: Chenglu Jin

CUSP Students: Shreeraman Arunachalam Karikalan, Aparna Bhutani, Siqi Huang, and Vivek Patel

Project Website: https://trajectorycapstone.wixsite.com/capsite

Recent years have seen an alarming increase in the usage of digital devices with location capabilities in them. Despite being the most sensitive data of individuals, mobility/trajectory data are collected by various devices, such as, smartphones and cameras. With pervasive usage of these personal devices every user leaves a non-erasable digital trace and is prone to being exposed to privacy risks. This research tries to quantify the privacy leakage in a given mobility dataset after it has been anonymized with a location privacy protection mechanism (LPPM). By privacy, herein, it means identity of a particular user. The research has been conducted from two different roles, publisher and adversary. As a publisher we anonymize the dataset using different anonymization schemes, while from the adversarial approach we try to recover and reconstruct the anonymized dataset. The research tries to quantify the privacy leakage based on how much of original trace is being recovered by the adversary. In addition, the research also automates the trace generation process for different trajectory data sets. Our research will be useful to increase privacy of the mobility data being collected everyday and making it difficult for attackers to obtain sensitive information of the users.

Digital CEQR 2.0: Real-Time Prediction of City Planning Proposals’ Environmental Impact

Sponsor: inCitu (by DRAW Brooklyn LLC)

CUSP Mentor: Chenglu Jin

CUSP Students: Guilherme Louzada, Chenjie Su, Akash Yadav, and Eric Zhuang

Project Website: https://ericzhuang0.wixsite.com/website-1

Project Tool Website: https://nyucusp-2020-ceqr.herokuapp.com/

Planning in NYC is mostly based on neighborhood initiatives and private land-use projects. Among those projects, many of them require the changes of the initial land use classification, for example, an industrial warehouse being retrofitted as a residential building. This process requires a multidisciplinary approach, including environmental and socioeconomic analysis, which is performed by the Department of City Planning, to make sure there will not be a significant adverse environmental impact on the neighborhood. The analysis is typically conducted under the scope of CEQR, which stands for city environmental quality review. The objective of this project is to develop a solution that will enhance the process involved in the generation of the environmental and socioeconomic information needed for this process, making use of machine learning techniques aiming to make the process quicker and more accurate.

Mapping Sustainable Mobility in NYC Nightlife Culture

Sponsor: VibeLab

CUSP Mentor: Kim Mahler

CUSP Students: Nicholas LiCalzi, Kaifu Ren, Yingyuan Zhang, and Yutong Zhu

Project Website: https://valeriezyy.wixsite.com/capstone

Our research aims to capture the vibrancy and expansiveness of New York City’s (NYC) nightlife landscape by interrogating how people access it: with Citibike, subways, or for-hire vehicles (FHV). We establish an understanding of the night-time transportation landscape to enable targeted recommendations around (increasing) transportation usage and mode choice that can ensure the continued survival of nightlife establishments throughout the city, the employment of people who work in them, and their accessibility and importance to their communities and patrons. Our initial research goal was to use multi-modal time series transportation data as an exploratory and advocacy tool to help local stakeholders in the nightlife industry (dance clubs, music venues, bars, and more) demonstrate their contribution to NYC’s citywide and hyper-local night-time economies. With the arrival of the COVID-19 pandemic, we expanded our research to examine how the night economy has been uniquely impacted by the crisis in the hopes of enabling policymakers to develop a nuanced understanding of the role of nightlife in the City and develop a uniquely targeted package of policies and aid that ensure its continued vitality. We built a baseline model of winter day and nighttime transportation with data from 2019, and expect to see significant variation given the arrival of the COVID-19 pandemic and related disruptions. We posit that the disruptions have severely affected night-life venues given their in-person nature and expect to see that reflected in the travel data for nightlife hotspots versus in the city as a whole. To empower future research we are sharing the complete dataset, consisting of aggregated venue data, travel data by mode, and more. Data analysts, city planners, venue owners, and others should be able to build on our work and better advocate for the needs of crucial players in NYC’s urban nightscape.

Preserving NYC Nightlife Culture Post-COVID-19: Keeping the "City that Never Sleeps" Awake

Sponsor: Office of Nightlife, NYC Mayor’s Office of Media and Entertainment

CUSP Mentor: Kim Mahler

CUSP Students: Vivian Chen, Ani Harnur, Andrew Norris, and Christine Vandevoorde

Project Website: https://nycnightliferesilience.github.io/

New York City nightlife is a bustling $35.1 billion industry with world-renowned culture and community that draws people from all over the world. With rising rent costs, gentrification, and most recently, Covid-19 shutdowns, the NYC Mayor’s Office of Media and Entertainment (MOME) is exploring supportive policies that can help the nightlife industry recover, adapt, and thrive for years to come. By analyzing MOME’s Covid-19 survey and incorporating neighborhood factors known to affect the well-being of the nightlife industry, the team will develop an interactive dashboard that displays the neighborhoods predicted to be at highest risk of experiencing permanent nightlife cultural loss. Additionally, text analysis of widely used sources describing local nightlife will reveal each neighborhood’s defining cultural characteristics, thus contextualizing and humanizing potential losses.

Documenting the Economic Impacts of New York City’s Investment in Water Supply: A Data Visualization Foundation for Economic Analysis

Sponsor: New York City Department of Environmental Protection

CUSP Mentor: Huy T. Vo

CUSP Students: Asnat Ghebremedhin, Angelia Lau, Ross MacWhinney, and Pratik Watwani

Project Website: https://pratikwatwani.github.io/cuspcapstone/

The New York City (NYC) Water Supply System, consisting of the Catskill/Delaware and Croton Water systems in upstate New York, provides one billion gallons of potable water to New York City’s 8.5 million residents every day. With an investment approaching $3 billion from 1993 to 2019, NYC Department of Environmental Protection (DEP) is interested in understanding the economic impacts of source water protection. This project develops an interactive, web-based visualization tool for exploring trends in and relationships between DEP’s Land Acquisition Program and economic metrics including annual Real Average Salary, Average Employment and number of Establishments in NYC’s watersheds. A literature review in the field of ecosystem service valuation was additionally developed with key interest from DEP.

Mapping Economic Change in New York City’s Telecom Industry

Sponsor: New York City Department of Information Technology & Telecommunications and Town+Gown

CUSP Mentor: Junaid Khan

CUSP Students: Wesley Chioh, Yichen Liu, Erik Lopez, and Ziyu Yan

Project Website: https://nyconnect2020.github.io/

The New York City Department of Information Technology and Telecommunications (DoITT) wants to understand if the current telecommunication industry regulations are effective at generating a competitive industry environment that can provide cheap and reliable internet access to all New Yorkers. Understanding this industry is important for the renewal of franchise agreements between NYC and internet providers in July 2020, which will have a lasting impact on the long-term economic benefits related to broadband internet access. This project tested for the presence of an optimal number of Internet Service Providers (ISPs) in New York City that can improve Internet accessibility. Current internet infrastructure coverage data was overlaid with pricing and socio-economic data to determine the relationships that can lead to the desired state. Tweets involving telecom were mined to analyze New Yorkers’ sentiments towards ISPs and the broader market, thereby reflecting the status quo and guiding the relevant policy interventions to address the raised concerns. It is estimated that there is an optimal number of ISPs for the NYC urban internet market but is yet to be achieved due to infrastructural and socio-economic unevenness. Furthermore, NYC’s government should enhance connectivity amongst less well-connected demographic groups through means-tested subsidies.

Assessing the Circular Economy Opportunity in NYC

Sponsor: New York City Economic Development Corporation (NYCEDC) and New York City Department of Sanitation (DSNY)

CUSP Mentor: Martina Balestra

CUSP Students: Yong Fan, Yang Li and Rongjian Yang

Project Website: https://yl1205.github.io/

This capstone project studied the circular economy concept mainly focused on understanding and predicting waste generation patterns from a residential waste perspective in the normal period approach and the Covid-19 period approach. Technical models including LSTM, Cluster, and Regression models were applied in the two major analyses to see the trend and predict the waste generation volume for each waste type as well as understand how socioeconomic factors would affect waste generation.

Smart Monitor for Accelerating Regional Transformation (SMART)

Sponsor: US Ignite

CUSP Mentor: Martina Balestra

CUSP Students:Jianqi Tang, Ram Sowmya Narayanan, Yanyan Xu,  Zehui Xiang, and Zheyuan Zhang

Project Website: https://smartcapstone.wixsite.com/nyucusp

Many cities have identified gaps between big data, Internet-Of-Things (IOT) and the potential of using it for economic development decisions. Unlike multi-national corporations, small businesses lack the capacity, resources, and budget to conduct research and compare the pros and cons of locations. The purpose of this capstone is to develop a Decision Support System (a recommendation system), including a smart framework and an online interactive tool, for small businesses, using the City of Portland, Oregon to test our product. The system will highlight, and rank locations based on their economic and demographic information to help small business owners make better decisions using a scoring mechanism similar to that used by multinational corporations. We expect this platform to empower small businesses in their decision-making process and city officials to get an idea of how well their city is supporting small businesses.

Applying Multi-Agent RL to SLAM with Graph Pose for Sampled-Data MPC and CPN of Autonomous Drone Swarms

Sponsor: RiskEconR Lab for Decision Metrics @ Courant Institute of Mathematical Sciences NYU (RiskEconR Lab @ Courant NYU) with Agile Robotics and Perception Lab (ARPL) @ Tandon School of Engineering

CUSP Mentors: David K A Mordecai, PhD and Giuseppe Loianno, PhD

CUSP Students: Rohith Gandhi Ganesan, Yue Jin,  Kunru Lu, and Elmon Toraman

Project Website: https://yuejinyz.github.io/2020CapstoneProject.github.io/index.html

Developing methods that allow drones to autonomously navigate in different environments has been a topic of extensive research in recent years. One research topic of interest for interest in autonomous drone navigation is to explore the maneuverability and capability of drones to navigate inaccessible environments and situations that might be too risky for human access. Using a swarm of drones that autonomously navigate a postcatastrophe scenario in order to optimally map the disaster zone, i.e independently and efficiently identify and map the structural damage across a geographic site, has been a problem less explored. Detection and mapping changes across a post-catastrophe site enables a more robust estimation of structural damage. This project1 attempted to explore and simulate a reinforcement learning approach to enable drones to perform task assignment and scheduling in order to efficiently maximize coverage for identifying and mapping structural changes within the post-catastrophe environment. The primary objective of the simulation was to focus on the exploration of ad-hoc decentralized task assignment and scheduling by one or more drone(s) at the edge with minimal connectivity aside from local communication between nearest neighbors. Other workstreams in the project explored satellite and aerial imagery, seismic structural damage equation models and generative adversarial networks (GANs) related to the Port-au-Prince 2010 Haiti earthquake site as a use case and attempt to explore methods that might be utilized to identify structural changes from satellite images, using generative synthetic data and estimated fragility equations in order to address uncertainty and ambiguity in the detection of discrepancies in edges related to damage which could aid the drones in the uncertain areas.

CIV-LAB COVID Team

Sponsor: CIV-LAB

CUSP Mentor: Stanislav Sobolevsky

CUSP Students: Earl Lin, Jerome Louison, Yushu Rao, and Xinran Zhao

Project Website: https://xinran46.github.io/

Github Page: https://github.com/ImpactsOfCOVID19InNYCCommunity/NYU-CUSP-Capstone-2020

The goal of this project is to leverage real-time urban data to help New York City government and community organizations identify community-level response to the Covid-19 epidemic within a low-income neighborhood disproportionately affected by the disease, using Brownsville, Brooklyn as the study case. We determined that in fact there was a statistically significant difference in the use of government services (including a 7% drop in 311 calls, and 51% decrease in MTA ridership) as a result of the “stay-at-home” order. Additionally, we identify the Brooklyn community districts (CD) and zip codes most similar to Brownsville (using a cosine similarity matrix), and the “informal” attitudes of Brownsville residents (by sentiment analysis of Twitter data). Lastly, we conducted OLS regressions to determine the correlation between the dataset variables with cases of Covid-19. All regressions conducted showed a high level (R2 > 0.84) of explanatory power with relation to Covid-19 related variables, and the correlation between independent and dependent variables was statistically significant (past p=0.05 threshold) for most factors. Deliverables included a data visualization tool (ie: webpage) of results, and one-page summary memo. Our methodology could potentially serve as a blueprint for low-cost, real-time analysis to inform public policy for Brownsville, as well as similarly situated neighborhoods in NYC.

Crowdsourced Security Cameras Enabling a Real-time Scaled Response to Crime

Sponsor: City of Paterson, NJ

CUSP Mentor: Junaid Khan

CUSP Students: Jianan Gong and Pengcheng Yang

Project Website: https://63576694.wixsite.com/capstone

The City of Paterson plans to utilize crowdsourced security cameras to reduce the crime rate by getting a real-time response to the crime and recognize the social distancing violation in this COVID-19 situation. By analyzing geospatial data, the team plots the overlapped heatmap of both crime and COVID-19 response and finds the best places to install cameras. The further clustering of temporal and spatial patterns of crime helps the city to better monitor violent crimes at midnight and theft in the early morning, while the time prediction for each census tract provides the police key regulatory direction in the future.  After the installation of cameras, the team develops a dashboard of public surveillance cameras for social distance detection by using YOLO4, the object detection algorithm.

Interagency Work Zone Traffic Data Modeling and Analysis

Sponsor: HDR

CUSP Mentor: Kaan Ozbay

CUSP Students: John Collier, Seunggyun Han, Linda Jaber, and Akhil Singh

Project Website: https://workzone-collision-analysis.github.io/

Road construction events are a necessary part of keeping road infrastructure in good condition but can pose significant safety problems when implemented. Transportation authorities in NYC seek a better understanding of the type, severity, and extent of mobility impacts associated with work zones. This project proposes using a k-means clustering approach to predict the probability of a vehicle collision occurring in the proximity of a road construction event (i.e work zone). The proposed clustering method is applied to over 20,000 construction and emergency construction events of relatively short duration in New York City to identify types of work zones that may present greater safety risks. This methodology builds upon the existing body of research by utilizing only publicly available datasets and by applying the methodology to roads and highways in the five boroughs of New York City. The results of this project are in service of enabling practitioners to employ appropriate mitigation strategies during project programming, design, and in the development of effective transportation management plans. This project is part of the Center for Urban Science and Progress capstone process with capstone sponsor HDR and a consortium of transportation authorities in the New York City Metro Area.

Urban Dynamics of Bird Migration

Sponsor: Cornell Lab of Ornithology

CUSP Mentor: Charlie Mydlarz

CUSP Students: Max Brueckner-Humphreys, Mei Guan, Martha Norrick, and Xin Yu

Project Website: https://www.soundsofspring-cusp2020.com

Avifauna are among the most prominent animals urban dwellers see on a daily basis. They are critical to urban biodiversity but their numbers have declined rapidly over the last 40 years (Rosenberg et al., 2019). However, emerging methods and technology such as acoustic sensing, complex acoustic DNN (Deep Neural Net) models (YAMNet & BirdVoxDetect), NEXRAD radar data, hourly weather data, and real-time citizen science data (eBird) offer researchers an opportunity to better understand bird populations and migration patterns along the Atlantic Flyway during peak spring migration. Our project focuses on consolidating these data sources, understanding their relationship with one another, and establishing the efficacy of such techniques for studying urban avifauna around Washington Square Park during spring migration at the dawn chorus hours (4am-8am EST).

As the evolving phenomena of COVID-19 takes its toll on urban life, we found that SPL (Sound Pressure Level) data around Washington Square Park (WSQ-NYC) was the lowest it has been across the last 4 years (Appendix Figure 2) offering researchers an unprecedented time to research the acoustic environment of the park. When testing the efficacy of the BirdVox Detect model in the urban soundscape, we found that both precision and recall were markedly low – the BirdVoxDetect model does not perform well in the WSQ-NYC area. Using a multivariate regression model with decomposed trend values, we found no correlation between the number of positive bird classifications by YAMNet and the migratory traffic rate (MTR) as detected by NEXRAD radar. However, we found a weak correlation between avian acoustic energy and anthropogenic noise as recorded by SONYC sensors.