Tag Archives: Spatial Anonymization

24 Feb 2021

Spatial Anonymization in Household Surveys

Location is an important factor in population and health outcomes. Knowing the geospatial location of household survey clusters allows researchers to analyze the impact of location on peoples’ health, nutrition, and access to health care services. Geospatial data provide a clearer picture of where progress towards the Sustainable Development Goals is and is not being made at subnational levels.

The United Nations’ Inter-Secretariat Working Group on Household Surveys (ISWGHS) was convened to identify priority areas for household surveys to meet new data demands and increase their impact on policy and research over the next decade of the 2030 Agenda for Sustainable Development. Under the ISWGHS, The DHS Program and the World Bank co-led a task force on Spatial Anonymization of Public-Use Household Survey Datasets.

The DHS Program has collected GPS coordinates for household survey clusters since 1996. To ensure respondent confidentiality and prevent positive identification (disclosure) of respondent locations, the GPS position of each urban cluster is displaced by up to two kilometers and up to 5-10 kilometers for rural clusters. This method of geomasking coordinates developed by The DHS Program is straightforward and has been widely accepted by analysts using DHS geospatial data. Nonetheless, there are legitimate concerns that urban points may be overly displaced, reducing the analytical usefulness of the geospatial data, and that some rural points may not be adequately displaced to ensure respondent confidentiality. In response, the spatial anonymization task force convened to explore more sophisticated methods of anonymizing geospatial data.

The task force developed and tested new population-based displacement tools on multiple DHS survey datasets. These tools use an area’s population to determine the minimum distance a cluster’s GPS position must be displaced. These new methods show promise over current spatial anonymization methods to better protect survey respondents while minimizing any adverse impact on analysis and continue to be explored using DHS datasets.

The task force also outlines immediate steps that can be taken to protect respondents. “Even without switching to a new population-based approach [to anonymize geospatial data], we should take steps to verify that we are within an acceptable level of disclosure risk and that our current anonymization objectives are being achieved,” explains Trinadh Dontamsetti, Lead, Geospatial Research. Standards in data protection and security have evolved—the European Union General Data Protection Regulation requires that personal data, including location data, be safeguarded. The task force recommends assessing the risk of disclosure. By quantifying and measuring spatial disclosure risk, the risk can be managed.

Read the “Spatial Anonymization” report for more from the Spatial Anonymization of Public-Use Household Survey Datasets task force. This report was also discussed by stakeholders to prepare a forthcoming paper on “Positioning Household Surveys for the Next Decade.” Both papers will be presented as part of the 52nd Session of the United Nations Statistical Commission: Better Data Better Lives on March 1-3 and 5, 2021.


18 Nov 2020

The DHS Program Geospatial Team Celebrates GIS Day

November 18, 2020, is GIS Day, a day to celebrate Geographic Information Systems (GIS) technology and their application to DHS data. This GIS Day, we highlight recent work from The DHS Program Geospatial team.

Highlighting Subnational Inequalities That May Hinder COVID-19 Mitigation

The Geospatial team produced a story map using DHS spatial data on handwashing and the average number of people per sleeping room in some countries in Africa and Asia to illuminate subnational variations and inform COVID-19 mitigation strategies.

Health Data Mapping Workshop Facilitated Virtually

In February 2020, The DHS Program released a call for applications for a Health Data Mapping regional workshop, to be held in Hyderabad, Telangana, India in May. When this workshop was canceled due to the COVID-19 pandemic, the Geospatial team quickly pivoted and developed an online course with the same basic objectives. A new call for applications was launched targeting countries in East Africa. Of 175 applicants, 25 were selected according to established criteria, including 15 men and 10 women.

The redesigned course, a semi-synchronous workshop, was conducted from July to September and included self-paced eLearning and instructor-led virtual sessions. Each week, participants completed readings, videos, and assignments. Participants acquired knowledge on data cleaning, converting GPS data in Excel, and joining data in QGIS, an open-source GIS software package. For their capstone assignment, participants created and presented maps using data from DHS surveys.

The DHS Program periodically hosts virtual instructor-led courses like Health Data Mapping on The DHS Program Learning Hub. For future offerings of the Health Data Mapping and other courses, check The DHS Program website for calls for applications.

Two Spatial Analysis Reports Published

Predicting HIV/AIDS at Subnational Levels using DHS Covariates related to HIV uses GPS data collected in DHS surveys to produce spatially interpolated maps to predict indicator values at non-surveyed locations. This report produces fine spatial and lower level estimates of HIV prevalence for seven Global Fund-supported countries in sub-Saharan Africa. A multi-task Gaussian process is used to leverage HIV prevalence data collected in DHS surveys along with HIV-related indicators such as sexually transmitted infection (STI), condom use, and sex partners. The model produces estimates of HIV prevalence along with uncertainty at the second subnational administrative level (ADMIN 2) where health programs are designed and implemented. This geospatial modeling approach was first described in Interpolation of DHS Survey Data at Subnational Administrative Level 2, and summarized in a previous blog post.

Geospatial Covariates: Proxies for Mapping Urban-Related Indicators uses DHS data from three East African countries to explore what makes an enumeration area rural or urban. The DHS Program does not define urbanicity; countries determine official urban and rural classifications themselves. This can impair comparative analyses of the relationship between urbanicity and health outcomes. In this Spatial Analysis Report, researchers sought to determine if geospatial covariates of urbanicity, including intensity of nightlights and travel times to hospitals, can predict urban-correlated health and demographic indicators from DHS surveys.

Spatial Anonymization in Public-Use Household Survey Datasets

The DHS Program Geospatial team is responsible for displacing cluster location data for all DHS surveys to maintain the confidentiality of DHS survey respondents. This anonymized spatial data is publicly available on The DHS Program Spatial Data Repository to meet research needs. The DHS Program Geospatial team is collaborating with other institutions, such as the World Bank and the United Nations to develop a forthcoming set of guidelines on spatial anonymization of public use microdata.

The information provided on this Web site is not official U.S. Government information and does not represent the views or positions of the U.S. Agency for International Development or the U.S. Government.

The DHS Program, ICF
530 Gaither Road, Suite 500, Rockville, MD 20850
Tel: +1 (301) 407-6500 • Fax: +1 (301) 407-6501
dhsprogram.com