Improving Smart Heating Efficiency through Behavioral Energy Analytics

The Solution

Big Blue AI developed an end-to-end data intelligence pipeline and behavioral clustering framework for Tiko Energy Solutions, analyzing millions of telemetry records across 48,000 smart heating devices to maximize household energy savings while maintaining consumer comfort.

• Data Homogenization: Built robust data integration infrastructure to ingest and standardize 13 disparate database tables covering 48,000 electric heating devices across France, integrating geospatial API elevation data into a unified analytical dataset.
• Size-Independent Energy Metric: Engineered the innovative Savings-to-Power Ratio (expressed in hours), an advanced feature that decorrelated energy savings from house size to establish a qualitative, physical benchmark showing the exact daily hours heaters were turned off via the Tiko app. 
• Smart Household Behavioral Segmentation: Created an intelligent segmentation model that automatically categorizes households into distinct energy-saving profiles and isolates key behavioral patterns driving high consumption.
• Data-Driven Intervention Rules: Formulated rule-based optimization thresholds proving that functional scheduling, varying setpoints at least twice daily with temperature steps under 4°C, maximizes thermal efficiency.

The Challenge

Tiko Energy Solutions offers smart room temperature programming apps designed to achieve up to 20% in household energy savings. However, Tiko faced the complex challenge of evaluating 5 gigabytes of highly fragmented, multi-granularity telemetry collected from 48,000 customer devices in France over an 8-month winter window. Without a way to decorrelate property size from actual user efficiency or categorize diverse usage habits, Tiko lacked the actionable behavioral insights needed to guide low-saving users toward optimal heating habits without sacrificing indoor comfort.

The Impact

The implementation of this project unlocked immediate operational clarity and tangible commercial strategies for Tiko's product team:

• 48,000 Devices & 5GB Data Unified: Successfully cleansed and harmonized multi-source time-series telemetry across 48,000 smart heaters into a master analytical framework.
• 60% Savings Opportunities Identified: Pinpointed that while 60% of users currently maintain flat, static temperature schedules, migrating them to functional scheduling unlocks maximum energy efficiency.
• Quantified Optimal Usage Formula: Scientifically validated that executing at least two daily temperature adjustments with step changes below 4°C delivers the highest energy performance.
• Direct Pathway to Automated Engagement: Established the foundation for personalized monthly app notifications to directly drive user retention and energy reduction.

How Big Blue AI helped

Big Blue AI acted as the lead data science partner throughout this collaboration with tiko Energy Solutions. Our team architected the pipeline to clean and process 5GB of fragmented telemetry across 13 database tables and geospatial APIs. To overcome property-size bias, Big Blue AI engineered the Savings-to-Power Ratio, decorrelating energy efficiency from house size. Using machine learning clustering and statistical testing in Python across 48,000 active devices, we converted complex time-series data into actionable behavioral thresholds. This provided tiko with a scalable foundation for automated in-app notifications, user rewards, and predictive savings models.