Extreme Weather Impact
on Facility Maintenance

What we wanted to learn

To analyze how climate factors such as freeze-thaw cycles, precipitation, and Humidex influence the Facility Condition Index (FCI) of buildings in Hamilton. The project also predicts how these factors will evolve, so the city can plan maintenance budgets proactively rather than reactively.

Techniques applied

• Data Integration: Geospatial mapping and merging of climate and building datasets.
• Correlation Analysis: Scatter plots and correlation metrics to surface the strongest climate → FCI signals.
• Predictive Analytics: XGBoost and SPSS for predictive modeling.
• Scenario Testing: Sensitivity analysis of climate variables under multiple future scenarios.

Models used

• XGBoost: Non-linear predictive modeling with high accuracy.
• Statistical Regression: Exploring linear relationships between climate variables and FCI.

Poster, dashboard and assistant

The findings were materialized in three deliverables: an academic poster summarizing methodology and conclusions, an interactive dashboard for stakeholders to explore the data, and an LLM-powered assistant that answers questions in natural language about facility risk.