Alert: Please correct the errors!
GIS displays typically contain many data layers ranging in type and level of detail (LOD) often resulting in dense map displays suffering from clutter and occlusion. This can lead to cognitive overload requiring excessive manual interaction to tailor and filter the scene, usually done globally. The result is a loss of context due to changes in LOD, map region/scale, and display content and often involves refresh delays. Through an IR&D project at MITRE, a non-profit organization chartered to work in the public interest, we developed a plug-in to ArcMap to investigate a localized, "detail on-demand" filtering strategy that may be a more efficient and desirable approach than global filtering. This enables the creation of semantic lenses that act as light-weight customizable visual filters. The current plug-in provides basic layer filtering (on/off), with near-term enhancements to provide other filters such as LOD (view/scale dependent content), magnification, and query-based filters.