Geospatial footprint library of geoparsed text from geocrowdsourcing

Abstract

AOA: The research paper reports on the generation of geospatial footprints from geoparsed text associated with geocrowdsourced platial data collected and stored in the George Mason University Geocrowdsourcing Testbed (GMU-GcT). The GMU-GcT facilitates study of social dynamics, quality assessment, data contribution patterns, and position validation for geocrowdsourced spatial data, with a primary purpose of mapping transient obstacles and navigation hazards in a dynamic urban environment. This paper reports on the automated generation of geospatial footprints using open-source software, and discusses the role of automated geospatial footprints in quality assessment for automated position validation. A detailed, local gazetteer is used to store placenames and placename variants including abbreviated, slang, former, and jargon-based instances. Obstacle reports containing location descriptions are geoparsed and processed with the help of the GMU-GcT gazetteer to generate geospatial footprints, which are used in a quality assessment process to validate the position of obstacle reports. Continuing research with the GMU-GcT has produced fifteen characteristic footprints types, which are generated and grouped into simple, complex, and ambiguous categories. The open-source tools used for generating these general footprints include MapBox, TURF.js, jQuery, and Bootstrap.