End User
City of Athens
Description
The Building Energy Consumption Anomaly Detection service identifies unusual or unexpected patterns in municipal building energy usage by analyzing historical consumption data. Using machine learning models, it detects sudden spikes, drops, or gradual deviations from expected behavior. This service enables municipalities to quickly uncover equipment malfunctions, energy leaks, operational inefficiencies, or billing errors, ensuring faster corrective actions. It supports preventive maintenance, reduces unnecessary energy costs, and improves overall building performance monitoring.
Core Capabilities
Monitoring & Anomaly Detection
Optimization & Decision Support
Business Need
It detects unnoticed system failures or equipment malfunctions early before they cause high costs, identifies abnormal energy consumption patterns such as leaks, operational inefficiencies, or user behavior changes, and supports proactive maintenance and energy management strategies.
Key Performance Indicators
Precision, Recall, F1, False Positive Rate
Data Provided
Building Metadata (size, building category, year of construction) - 500 buildings with the potential for expansion -no documentation-static data
Historical Energy Consumption Bills – 500 buildings – no documentation – monthly
Historical Energy Consumption – 50 buildings with the potential to add more as smart meter installations progress – no documentation – hourly
External datasets (calendar days)
TEF
TEF BUILD