Dissertations, Theses, and Capstone Projects

Date of Degree

9-2026

Document Type

Master's Thesis

Degree Name

Master of Science

Program

Data Analysis & Visualization

Advisor

Brett F. Branco

Subject Categories

Applied Statistics | Data Science | Environmental Monitoring | Hydrology | Longitudinal Data Analysis and Time Series | Multivariate Analysis | Oceanography

Keywords

Wastewater Treatment Plant (WWTP), Nitrification, Biological Nitrogen Removal, PCA Biplot, Machine Learning

Abstract

Jamaica Bay, located along the southeastern coast of New York City, acts as a biodiverse estuary of wetlands, meadows, and salt marsh islands. The purpose of this study is to analyze the water quality conditions of the region over time, comparing locations around the bay to identify hyperlocal features that influence larger trends in the hydrological system. Ten variables were used as water quality indicators, including total Kjeldahl nitrogen, salinity, pH, Secchi disk depth, and total phosphorus, among others, across five stations in the bay, between 1994 and 2024. After data cleaning and standardization methods were applied, principal component analysis was implemented, and k-means clustering was applied to the data for trend identification. Across all five stations, a major changepoint was detected by the clustering around the years 2014 to 2016. Additionally, 2002 emerged as an outlying point in all PCA analyses, warranting further research and explanation. An analysis of the loading contributions in the PCA revealed that biological and nutrient-based features were the primary driving factors of explained variance at stations located inside the bay, with physical and hydrological features overtaking this explained variance as water flows out of the bay.

Additional Files
Jamaica-Bay-Water-Quality_Capstone-main.zip (30922 kB)
GitHub Repository
Download

Share

COinS