A Context-sensitive Approach to Anonymizing Spatial Surveillance Data: Impact on Outbreak Detection

Abstract

A population-density-based Gaussian spatial blurring markedly decreases the ability to identify individuals in a data set while only slightly decreasing the performance of a standardly used outbreak detection tool. These findings suggest new approaches to anonymizing data for spatial epidemiology and surveillance.

UK open source crime data: accuracy and possibilities for research

“…Such techniques aim to minimize the extent to which the anonymized data can be "reverse geocoded" -which would threaten an individual's privacywhilst preserving the general spatial pattern in the data. Areal aggregation is one technique but this can lead to a loss of information (e.g., Wieland et al 2008) and so spatial epidemiologists have devised a number of alternative methods using affine and randomizing transformations to name a few (e.g., Armstrong, Rushton, and Zimmerman 1999;Cassa et al 2006;Wieland et al 2008). In terms of the impact of geomasking techniques, Leitner and Curtis (2004) examined how different methods affected student's perceptions of the spatial pattern of a sample of Homicide victim's residences in Baton Rouge.…”

“…First, census units are (approximately) standardized by population, and consequently the geographic size of a census area provides an indication of "urban density" or how much open space there is likely to be within it. As larger areas will have lower urban density (for a discussion of geomasking techniques based on population density, see Cassa et al 2006), they are likely to have fewer snap points within them, and hence:…”

Defining a Threshold Value for Maximum Spatial Information Loss of Masked Geo-Data

“…The methods that they used included visualization of point patterns, visualization of 2D and 3D density surfaces, examination of maps of density differences and cross K function analysis. Additionally, other scholars used spatial statistics to quantify the effects and, in particular, cluster indices, such as the clusters' sensitivity, specificity, detection rate, accuracy and the most significant cluster [23][24][25][26].…”

Wider Dissemination of Household Travel Survey Data Using Geographical Perturbation Methods

“…Kwan, Casas, and Schmitz (2004) distinguish three subclasses of jittered data points in which the masked location is either randomly located along or inside the perimeter of a circle with a center at the original location and a chosen radius, or the masked location lies randomly within any other polygon defined relative to the original point. Such jittering methods may be employed by taking into account the population density surrounding the respondent's location by skewing her address by a random offset based on a Gaussian distribution whose standard deviations are inversely correlated to the population density (Cassa et al, 2006). This approach displaces individuals residing in an area with a lower population density by larger amounts than their counterparts residing in urban areas characterized by a higher population density since the former individuals have a heightened disclosure risk (Rushton et al, 2008).…”

“…Thus, any statistical reflection of anonymity in a geographic perturbation technique must be directly linked to the number of additional individuals or households found within the exclusion area of the donut. By utilizing the k-anonymity concept introduced by Sweeney (2002), where k is a value representing the total number of households, the original household location cannot be reversely identified (Cassa et al, 2006). Allshouse et al (2010) provided the following operationalization of this concept of anonymity, which may be rearranged to reflect an inner-ring radius for the donut masking technique informed by the selection of a k-anonymity level by the data custodian.…”