Current trends in travel forecasting are motivating model updates which eschew the traditional zonal-aggregate demand components in favor of more disaggregate demand models. This generation of travel forecasting models places different requirements on the model user: a) model operation may span multiple software components, b) network and demand visualization may occur in disparate systems, c) user error is more costly with increased computational time and d) disaggregate ‘big data’ sets require new analysis tools. Almost all such projects strive for some type of model dashboard to tie these elements together and to increase model usability and transparency.

Here we propose one such approach, developed in collaboration with Metro Portland, which offers an interactive visual computing environment for model specification, operation, analysis, visualization and collaboration. The system incorporates components of the increasingly popular and open-source scientific Python ‘SciPy’ stack (tools which should be of increasing interest to the general planning community) and the DASH tour-based demand model together with network modeling components and services in Emme transportation forecasting software.

We will demonstrate how the framework supports model users in an interactive, visual computing environment that combines model steps, analysis, plots and rich media. We will show how the system illustrates the structure of the network models, the progress of model execution, and how it combines tools for ad hoc and automated analysis and visualization. We will explain how the system responds to user input and can how it can be used for effective collaboration.