The need for model integration arises from the recognition that both transportation decision making and the tools supporting it continue to increase in complexity. Many strategies that agencies evaluate require tools that are sensitive to supply and demand at local and regional levels. This in turn requires the use and integration of analysis tools across multiple resolutions. However, despite this need many integrated modeling practices remain ad hoc and inefficient.

A concept for an open-source data hub was developed to better enable the exchange of model information across multiple resolutions. All modeling and field data are fed and stored using a unified data schema. Tools within the data hub aid users in modifying modeling network, control, and demand data to match an objective, such as calibrating to field data. Visualization tools are built into the data hub’s core visualization program, NEXTA, along with powerful links to common web-based tools such as Google Earth, Google Maps, and Google Fusion Tables. The data hub reduces barriers to interfacing models across multiple resolutions and software platforms, which ultimately saves time and cost.

The following is a simplified, summary list of the most beneficial components of the data hub:

- Quick transfer of most common travel demand models into a mesoscopic DTA network

- Tools to adjust network, demand, and signal timing to facilitate quality DTA evaluations

- Linkages with signal timing optimization tools

- Linkages with microsimulation tools

Test applications were performed using the open-source data hub for two networks: an arterial network in Portland, Oregon and a freeway network in Tucson, Arizona. The data hub functionality was demonstrated by taking an existing regional travel demand model, exporting to DTA for mesoscopic analysis, exporting to a signal timing optimization tool, and lastly exporting to a microscopic simulation tool for detailed operations analysis. In the test applications each step and data conversion function are performed using the data hub and all model data are stored according to the data schema. Initial findings show that the data hub concept overcomes many of the previous shortcomings associated with integrated modeling applications.