Assessment of Volunteered Geographic Information Data Quality in The National Map Corps Project of the U.S. Geological Survey (USGS)

Abstract

In an effort to provide accurate and authoritative spatial data for The National Map of the U.S. Geological Survey (USGS) National Geospatial Program, the USGS began implementation of a new crowdsourcing project in 2010. The National Map Corps (TNMCorps) enlists volunteers to update structures data across all 50 states as well as Puerto Rico and the U.S. Virgin Islands. Volunteers collect and improve structures data by adding new features, removing obsolete points, and correcting existing data using a web-based mapping interface. Newly collected and modified point features become part of the USGS National Structures Dataset (NSD), a part of The National Map, which supplies data to US Topo maps, USGS cached base maps, and other derived products and services. Concern over the ability of volunteers to deliver high-quality data instigated a data quality study in 2012 during the Colorado pilot project, and a second nationwide quality study in July 2014. These data quality studies explore the quality of volunteered geographic information (VGI) within TNMCorps by assessing horizontal positional errors, attribute errors, and errors of commission. Results of the studies conclude that the quality of volunteered geographic data is significantly higher than baseline data, and the hierarchical editing approach improves the data at each stage. The results of the quality studies validate the overall data collection model of the project, and affirm that volunteers can provide high-quality geographic data with limited USGS monitoring.