Biography

Nthenya Kyatha is a PhD candidate in Computer Science at UMass Amherst, with a strong background in Electrical Power Systems Engineering. Her research focuses on utilizing low-cost IoT sensors and machine learning technology. This includes creating air-quality monitoring networks for urban transit, using vision technology to count passengers, and predicting electricity demand in cities, particularly as the need for cooling increases. Nthenya is involved in developing embedded firmware, conducting dynamic building simulations, and applying deep learning techniques to promote sustainable transportation and energy resilience. She is passionate about mentoring undergraduates and conducting research that benefits communities, striving to translate data-driven insights into equitable policies and resilient infrastructure. Outside of her academic pursuits, Nthenya is an avid hiker and runner who enjoys exploring local trails and participating in long-distance races.

Academic Status

PhD Student - 4th

Research Area/Department

Computer Science; Data Science; Engineering; Machine Learning/AI

Major/Specialty

Computer Science- AI specialisation

Degrees Earned or in Progress

BEng (honours) Electrical Power Systems Engineering, earned in 2020 PhD(Computer Science): in progress

Academic Preparation

Computer Vision, Machine Learning, Neural Networks, Artificial Intelligence, Distributed and Operating Systems, Algorithms, Networked Embedded Systems Design, Statistical Methods

Research/Publications

I am currently a PhD student at UMass Amherst, focusing on research in urban environmental sensing, electricity demand forecasting, and machine learning. Recently, I completed the design and laboratory validation of a dual-mode air quality monitoring network for bus stops and inside buses. This manuscript is now under review for publication, and plans for deployment of the sensor in Nairobi and Kigali are underway. In addition, I am working on two ongoing projects for which manuscripts are in preparation: Vision-Based Passenger Counting and Demographics Analysis: This project involves a multi-camera CCTV pipeline that detects, tracks, and re-identifies riders to quantify boarding and alighting events while estimating age and gender profiles. Electricity Demand Forecasting and Cooling Load Impact in Nairobi: This city-scale study combines satellite-derived building archetypes with dynamic thermal simulations to model residential cooling demand and assess its effects on transformer loading.

Research/Academic Interests

My research and academic interests are at the intersection of sensing technologies, data-driven modeling, and machine learning, with a particular focus on urban sustainability and mobility: Computer Vision for Public Transportation I develop end-to-end pipelines and models to detect, track, and re-identify passengers using on-board and curbside cameras. My work focuses on using appearance, temporal logic, and geomatics fusion for accurate boarding and alighting counts, considering bus occupancy and demographic profiles. Additionally, I create models that analyze driver behavior through camera footage and geomatics data to understand how driving patterns relate to accidents or incidents. Energy Systems & Demand Forecasting Modeling residential cooling loads under changing climate conditions by combining satellite-derived building archetypes with dynamic thermal simulations (EnergyPlus, JEPlus). I study how increased air-conditioning demand stresses distribution transformers and integrate occupancy and market-adoption data to predict future grid impacts. Environmental Sensing & Air Quality Designing low-cost, dual-mode sensor networks (solar-powered at stops, vehicle-powered on buses) to monitor PM₂.₅ exposure in real time. Focused on sensor calibration, over-the-air updates, and turning pollution data into actionable insights for transit agencies. Creating end-to-end IoT solutions for off-grid deployments, ensuring autonomous operation, cloud sync, and efficient power management in resource-limited settings.

Computational and Data Science Areas

Applied Computer Science; Computer Science; Electrical, Electronic, and Information Engineering; Infrastructure and Instrumentation

Motivation

I’m drawn to Sustainable Research Pathways for its inclusive, cross-disciplinary mentorship and its dedication to building lasting professional networks that extend well beyond a single summer. As a first-generation PhD student at UMass Amherst working at the intersection of environmental sensing, urban mobility, and machine learning, I’ve benefited enormously from mentors who champion diverse perspectives, though I am yet to collaborate deeply with peers and leaders across mathematics, engineering, and computer science. I’m eager to both learn from and contribute to a community that values every voice and transforms individual research into shared platforms for impact. During the summer project, I am excited to work alongside leading experts to master new methodologies, enhance my skills in data pipelines and vision modeling, and understand how my contributions fit into larger NSF-supported initiatives. I am a fast learner, capable of quickly adapting to new tools and techniques. Additionally, I have developed strong time-management skills, which allow me to contribute meaningfully to any project. I am particularly eager to see how scalable, cloud-enabled workflows are designed and deployed, and to gain insight into the career paths of professionals whose expertise aligns with my own. Over the year-long program, I look forward to grant-writing workshops, panel discussions, and networking events that will accelerate my professional growth and foster enduring collaborations. In return, I will mentor undergraduates interested in research, co-lead hackathons, and organize outreach seminars, helping to cultivate the supportive ecosystem that has fueled my development. My goal is that, together, we’ll create resilient, data-driven solutions that benefit communities and ensure every voice can thrive.

Lightning Talk Title

Computer Vision, Load Forecasting for Urban Mobility and Grid Planning

Keywords (Maximum 20 words)

Computer Vision; Edge AI; Embedded IoT; Air Quality Monitoring; Energy Modelling; Cooling Load Forecasting; Load Analysis; Sustainable Urban Mobility