Balancing geo-privacy and spatial patterns in epidemiological studies

• Main Authors: Chien-Chou Chen, Jen-Hsiang Chuang, Da-Wei Wang, Chien-Min Wang, Bo-Cheng Lin, Ta-Chien Chan
• Format: Article
• Language: English
• Published: PAGEPress Publications 2017-11-01
• Series: Geospatial Health
• Subjects: Geo-privacy; Geo-masking; Spatial epidemiology; D statistics
• Online Access: http://geospatialhealth.net/index.php/gh/article/view/573
• ISSN: 1827-1987; 1970-7096

To balance the protection of geo-privacy and the accuracy of spatial patterns, we developed a geo-spatial tool (GeoMasker) intended to mask the residential locations of patients or cases in a geographic information system (GIS). To elucidate the effects of geo-masking parameters, we applied 2010 dengue epidemic data from Taiwan testing the tool’s performance in an empirical situation. The similarity of pre- and post-spatial patterns was measured by D statistics under a 95% confidence interval. In the empirical study, different magnitudes of anonymisation (estimated Kanonymity ≥10 and 100) were achieved and different degrees of agreement on the pre- and post-patterns were evaluated. The application is beneficial for public health workers and researchers when processing data with individuals’ spatial information.