Abstract

The increasing pace of urbanisation and consumption has increased the burden of waste generation in the world and it is a huge burden on the current waste management systems. The United Arab Emirates (UAE) is a region that is intensifying this challenge through the accelerated urbanization, high sustainability targets, and the necessity of effective recycling policies. The thesis that is being examined explores the ways in which the data analytics can be utilized to streamline waste management and recycling trends, emphasizing the enhancement of the collection process, forecasting the trends in the waste generation as well as facilitating the use of the data in making the decisions within the area of the municipality. This research is based on the CRISP-DM (Cross-Industry Standard Process of Data Mining) model which offers a methodical way of converting unprocessed municipal information into actionable information. Based on a synthesis of municipal waste collection reports, IoT-enabled smart bin sensors data, and Geographic Information System (GIS) data, the study implements both quantitative and spatial analytics in recognizing inefficiencies, the high-waste areas and the trends in recycling behaviors in the chosen urban districts within the UAE. The six main questions that guided the research were (1) in what way data integration with multi-source information can increase predictive accuracy, (2) what spatial and behaviour factors can affect the performance of recycling, (3) howmachine learning models could be used to optimize the collection routes, (4) how the CRISP-DM can be relevant to municipal waste analytics, (5) how the use of data-driven dashboards can benefit operational decisions, and (6)what practical benefits are possible through predictive scheduling. Python (Pandas, Scikit-learn, XGBoost) and RStudio were used to process and model the data and supported by QGIS to visualize the data spatially. The models showed that predictive analytics can increase the efficiency of routes by 25-30% and decrease the events of overflow by 20 which are consistent with other international studies of the same nature. Clustering analysis showed that population density, land use type, and recycling performance have significant correlation and time series forecasting can correctly represent the periods of highest waste generation. The results show that combining the information collected by IoT sensors with the municipal data can help municipalities transition to proactive waste management. A decision dashboard written in Streamlit was also built to mimic the real-time route optimization, overflow warning, and recycling rate, which showed how analytics could directly guide operational planning. Results of this study support the conclusion that data-driven waste management can play an important role in supporting the objectives of the Zero Waste 2030 of the UAE by minimizing operating expenses, improving sustainability, and increasing the satisfaction of citizens. Some proposals include the necessity of unified municipal data system, constant investment into IoT infrastructure, and incorporation of data on citizen engagement in future designs. Further studies must consider the future opportunities of applying deep learning to automated image-based waste classification and predictive frameworks in a rural and semi-urban setting. These attempts would also improve the smart, circular, and sustainable urban development of the UAE.

Library of Congress Subject Headings

Refuse and refuse disposal--United Arab Emirates--Data processing; Recycling (Waste, etc.)--United Arab Emirates--Data processing; Predictive analytics; Machine learning

Publication Date

12-2025

Document Type

Thesis

Student Type

Graduate

Degree Name

Professional Studies (MS)

Department, Program, or Center

Graduate Programs & Research

Advisor

Sanjay Modak

Advisor/Committee Member

Philippe Bouvier

Campus

RIT Dubai

Plan Codes

PROFST-MS

Download

Share

COinS