The Florida’s Turnpike Enterprise (FTE) has re-architected its statewide travel demand modeling platform, the Turnpike State Model (TSM - NextGen), by embedding the entire modeling workflow within a custom QGIS plugin interface. This transition represents a fundamental shift from traditional, fragmented modeling environments toward a fully integrated, open-source, and GIS-native ecosystem.
At the core of this innovation is the use of QGIS not just as a visualization tool, but as the primary user interface for model configuration, execution, and diagnostics. All stages of TSM from land use editing and network preparation to population synthesis, demand modeling, and assignment are initiated and completed within the QGIS environment. This eliminates the need for external software, file transfers, or manual data handling across disconnected platforms, significantly reducing operational complexity and risk of error.
By leveraging QGIS’s open architecture, TSM has achieved complete independence from proprietary software, aligning with public-sector goals of cost efficiency, transparency, and long-term sustainability. The plugin-based design allows direct interaction with GeoPackage, PostGIS based land use and network data, where users can modify socio-economic attributes (e.g., population, households, employment) using standard QGIS editing tools. These changes dynamically trigger appropriate model components (e.g., population synthesizer, demand model, route choice), all orchestrated through an intuitive graphical interface.
A key advancement is the integration of AI-assisted workflows within the QGIS plugin. Customized GPT-driven user guides are embedded directly into the interface, enabling context-sensitive assistance for each modeling step. These guides are dynamically linked to backend GitHub repositories, providing users with real-time access to source code, documentation, and model logic for each widget or module. This creates a self-service, intelligent help system that reduces onboarding time and enhances user confidence in navigating complex disaggregate models.
The open QGIS platform has also enabled rapid acceleration of development cycles. AI tools have been used to generate, debug, and refine PyQGIS and UI components, allowing iterative improvements that would traditionally take significantly longer. This has resulted in a modern, modular, and extensible interface that can evolve alongside emerging modeling needs.
Ultimately, the TSM QGIS interface transforms how analysts interact with large-scale travel demand models bringing transparency, efficiency, and usability to a domain historically constrained by technical barriers and proprietary dependencies. This presentation will demonstrate the architecture, workflows, and lessons learned from deploying a fully open-source, AI-enhanced modeling platform within a statewide agency.