AT A GLANCE
Developed advanced clustering models to optimize decarbonization plans, reducing manual analysis time by 60-80%, cutting operational costs by 15%, and enabling targeted carbon emission reductions of up to 25% in selected buildings with improved isolation and LED lighting installation. These solutions streamlined planning, improved scalability, and set the foundation for efficient carbon reduction strategies.
Client information
CHALLENGE
The client sought to optimize their decarbonization planning process by segmenting buildings into clusters based on physical characteristics. This would allow them to better target carbon reduction strategies, automate processes, and handle complex building data at scale.
SOLUTION
Developed a K-Prototypes clustering model to segment buildings based on features like height, perimeter, and area, generating 15 clusters tailored to British Columbia for targeted carbon reduction plans. A similarity scoring system using cosine similarity and Euclidean distance was implemented to group buildings with similar decarbonization needs. Automated pipelines in BigQuery and Google Cloud Scheduler were built to enable real-time updates and data transformation.
IMPACT
TOOLS
Ready to achieve measurable results? Tell us about your challenges, and we’ll show you how we can help
