Case Study: Edmonton Fire Rescue Services
Edmonton Fire Rescue Services (EFRS) built Spark, a behavioural intelligence program that joins its own incident data with EA’s DemoStats, CommunityHealth and SocialValues to understand the pressures behind serious fires.
31 fire stations
1.3K full-time employees
1.2M Edmonton residents
780 square kilometres served
Challenge
Hot spot maps show what happened and when, but not why. Smoking shows the gap: Edmontonians smoke about 1.8 million cigarettes a day, yet there are about 90 recorded smoking-material fires a year, so the risk can feel remote.
Solution
-

Under a new analytics vision from Fire Chief David Lazenby, EFRS built a causal-chain framework: behaviours and symptoms shape motivations, which lead to actions and outcomes.
-

EFRS incident data, decades of narrative reports mined with natural language processing, and EA’s DemoStats, CommunityHealth and SocialValues together light up the whole chain.
-

From an ignition source to community context
1. Framework:
The causal chain - Behaviours, symptoms, motivations, actions and the outcome.
2. Co-occurence:
Traits that travel together - Smoking fires recur with homelessness, negligence and impairment.
3. Spark clusters:
Patterns, not one cause - Recurring combinations, each with its own audience.
4. Community context:
Local drivers - North: living alone, financial concern, anxiety. Southeast: chronic pain, lone parents.
Results
What Spark has delivered so far
- Explorer map: Plus early co-occurrence studies for prevention teams.
- Frontline tools: Including support for the Community Detectors Program.
- Relay prototype: Helping the Office of Emergency Management respond to mass displacement.
- Next: risk models: Behaviour-based community risk models and vulnerability indexes.