Leveraging NASA and other Earth observation data sources, ModelLab enables the design, storage, and execution of geospatial models and algorithms.  These geospatial models can, for instance, be used for precision agriculture, to simulate flow patterns of tsunamis and floods, to forecast crime or seismic risk, to determine lines-of-sight for cell tower siting, to interpolate among field observations, to compare coastal erosion patterns over time, and to monitor the impact of forest fragmentation, among countless other applications.

These types of geospatial modeling capabilities have been accessible to trained geospatial analysis professionals for many years, but ModelLab aims to make them available to non-technical users in a broad range of public and private sector positions. A Model Gallery will be initially populated with geospatial models prepared by members of Azavea’s research team.  In particular, it will include modeling “recipes” prepared by Dr. C. Dana Tomlin, originator of Map Algebra.  Public models in this initial prototype will be freely available for download and use in external applications through a Creative Commons or similar open license and will be searchable by keyword and topic.

At the core of the application, is GeoTrellis, Azavea’s open source high performance geoprocessing engine.  Based on the Scala language and the Spark framework, GeoTrellis provides the ability to rapidly process geospatial data by distributing the processing across computing clusters.  The goal of the framework is to transform user interaction with geospatial data by bringing the power of geospatial analysis to real-time, interactive web and mobile applications.

The development of the ModelLab project is supported by the NASA Small Business Innovation Research (SBIR) program, Award Number NNX15CS06P.