Transit Shelter Risk Assessment Tool: A predictive analytics solution for transit shelter risk assessment

The Problem

Prior to the tool's development, the Edmonton Transit Service (ETS) lacked standard procedures, formal guidelines or objective data to support infrastructure decisions. Management of these assets suffered from systematic inconsistency, making it impossible to objectively determine when to remove, replace or relocate shelters facing recurring issues. Without a centralized analytics system, ETS could not identify broader patterns in vandalism, misuse and nuisance complaints, leaving maintenance teams and security personnel unable to prioritize problematic areas or proactively deploy resources where they were needed most.

The Solution

The Transit Shelter Risk Assessment Tool was established to provide a consistent approach to infrastructure management. The system utilizes contextual and predictive analytics—specifically Risk Terrain Modeling—to evaluate spatial risk across Edmonton's roughly 2,500 shelters and 7,500 stops. Implemented with open-source tools within the existing budget.

The Impact

The tool provides actionable insights for proactive risk management, enabling data-driven decisions regarding security needs, monitoring and infrastructure optimization.

Lessons Learned

The project demonstrated that predicting infrastructure risk requires looking beyond historical incident data (like past vandalism). Integrating spatial and environmental risk factors—such as proximity to liquor stores, schools or homeless shelters—vastly improved the tool's predictive capabilities.