STEAM — measuring the economy of sport

STEAM is how Sport Tourism Canada puts a number on the value of a sport event — from a regional swim meet to a national championship. The economic model and weights were developed by Signal49, built on standardized visitor-expenditure profiles and Statistics Canada data. My job was everything around that model: the backend it runs behind, the authentication that protects it, and the front-end members actually use.

Where I came in

Signal49 owns the calculation engine. I built the platform that turns it into a product — the Python backend the front-end talks to, a complete authentication and authorization system, and the Next.js application on top.

What I built

• A FastAPI (Python) backend sitting between the Next.js front-end and Signal49's model service — validating and normalizing every payload, mapping events to locations and coverage, and adapting the model's output into the shape the UI needs.
• A complete authentication system on Auth0 — JWT verification, and role-based authorization spanning site admins, organizations, teams, and per-project ownership, with full audit logging of every privileged action.
• Auth0 Management API integration that provisions and syncs members from WordPress into Auth0, so existing STC membership flows straight into the platform.
• A multi-tenant data model — organizations, teams, members, and roles — so the tool serves a national membership cleanly.
• The Next.js front-end, bilingual (EN/FR), for both the self-serve and validated (STEAMPRO) studies.
• A print-perfect report exporter — every economic-impact report renders to a 1:1, pixel-accurate printable PDF (and Excel), so what's on screen is exactly what comes out of the printer for funding proposals and board packages.
• Deployment on Azure — pipelines, blob storage, and containerized services.

Design

The visual concept came from a design vendor; I adapted and rebuilt it to work properly on desktop, where members actually do this work — tightening the data-entry flows, tables, and reporting views for real screens.

The result

A dense economic model — someone else's hard science — wrapped in a platform a volunteer event organizer can use and a professional analyst can trust. Secure, bilingual, multi-tenant, and built to run for a national membership.