HVAC systems are often maintained reactively, with faults detected only after performance degradation becomes visible. For Alert Labs, this creates a gap between continuous sensor data collection and actionable maintenance insight. This project develops Anomalyze, a hybrid fault detection and decision-support system that combines anomaly detection, supervised learning, and an interactive diagnostic dashboard. The solution supports early fault identification, system-level monitoring, and unit-level investigation. Evaluation through core-based validation, preprocessing analysis, and usability testing demonstrated improved interpretability and stronger support for operational decision-making.

Companies in the HVAC sector lack efficient remote detection of system abnormalities, relying on customer complaints, leading to inefficiencies and unprepared technicians. Alert Labs, a smart solutions manufacturer, seeks to integrate advanced analytics with proprietary sensor technology for predictive maintenance and improved diagnostics. This project involves anomaly detection, predictive maintenance, and the development of a user-friendly interface for technicians. By leveraging IoT, machine learning, and cloud computing, the aim is to enhance HVAC service quality and operational efficiency.

This report discusses the problem of limited battery capacity in the electric aviation industry. It highlights the potential benefits and impacts of optimizing electric plane batteries, including reducing carbon emissions, lowering operation costs, and improving access to remote communities. The report identifies key stakeholders, including Pipistrel, Transport Canada, Waterloo Institute for Sustainable Aeronautics (WISA), and Waterloo Wellington Flight Centre (WWFC). It analyzes the efforts made to address aviation carbon emissions, battery limitations, and decision-making for flight safety. The project objectives of improving data preparation, visualization, and optimizing battery management are outlined, emphasizing the need for accurate models and user-friendly interfaces.

This report presents a battery-health decision-support system for electric aircraft. The system converts noisy battery-management-system and aircraft telemetry into event-level battery records, estimates a condition-aware latent state of health (SOH), forecasts near-term degradation, and translates these outputs into operator-facing advisory guidance. Results show that the latent SOH pipeline substantially reduces implausible raw SOH variation and enables useful short-horizon forecasting for maintenance and flight-planning support. The final design is compatible with existing telemetry interfaces and demonstrates a practical framework for battery-health monitoring, degradation forecasting, and conservative operational decision support in electric aviation.