A technology is presented that allows users to explore scientific geospatial data sets using a simple web browser. This platform brings many of the features typically associated with desktop visualization software to computers and smartphones at the lab or at conferences or field campaigns with a minimal data transfer footprint to the client, thus facilitating access to public scientific data for users and in scenarios where traditional solutions are impractical. Unlike typical data portals on the web today, data files are accessed on-demand to produce custom and arbitrary visualizations through interactive user interface elements on dynamic maps

The software requires a standard LAMP stack and GRaDS to run, and is completely configurable to the needs of the project. Standard visualization plugins (i.e. cross sections, timeseries, etc.) can be easily modified and entirely new plugins can be added to the visualization engine. Data may be hosted from local repositories or from remote servers over the DAP protocol. Data discovery and publication to the site is accomplished through configuration files that define data holdings that can be produced statically or dynamically from an automated metadata scraping program or web crawler that indexes data holdings on scientific data servers.