Stay up-to-date with what we are doing

Data Artefact Study

The GovLab Releases New Study on Population Density Data and Its Uses

Posted on 21st of September 2021 by The Govlab

The GovLab Releases New Study on Population Density Data and Its Uses
The GovLab Releases New Study on Population Density Data and Its Uses

Today, The GovLab is releasing a new report examining data and information about population density that can help to create social and economic value. The report, “Where Is Everyone? The Importance of Population Density Data” is the first in a new series of Data Artefact Studies from The GovLab. The piece released today focuses in particular on Facebook’s Population Density Map and several of its implementations to understand promising practices, opportunities, challenges, and risks in the use of population density data more broadly. 

The report begins with an exploration of new and traditional approaches to measuring population density, and ways in which density information has frequently been used by humanitarian, private-sector and government actors to, for instance, support long-term infrastructure development projects, plan vaccination campaigns, and determine geographic areas most likely to be impacted by climate change. It examines how several new innovations are leading to fresh ways of collecting data—and fresh forms of data—and how this may open up new avenues for using density information in a variety of contexts, when handled responsibly. 

The bulk of the study focuses on Facebook’s High-Resolution Population Density Maps (also referred to as HRSL, or high resolution settlement layer). HRSL has been deployed by research institutions, non-profits, and international development agencies to support a range of projects. The density maps, which estimate population levels within 30-meter grids in nearly every country in the world, are not based on data collected through Facebook’s social media or messaging platforms, but instead use a combination of machine vision AI, satellite imagery (aerial data), and census information compiled by Facebook’s AI and data science team in collaboration with Columbia University. The resulting density maps are publicly available through the Humanitarian Data Exchange (HDX) data store and the Registry of Open Data on Amazon Web Services (AWS).

In addition to an overview of the HRSL itself, the Data Artefact Study focuses in particular on a series of implementations of the HRSL with a brief discussion of each use case’s objectives, outcomes to date, lessons learned, and planned next steps. The case studies were developed through desk research and interviews with stakeholders involved in the use of HRSL, as well as Facebook personnel involved in its development and dissemination. The case studies are: 

  • Pulse Lab Jakarta’s efforts to identify areas with high transmission risk and transmission potential for the spread of Covid-19 in West Java, Indonesia: In September 2020, Indonesia registered the second highest transmission rate of Covid-19 cases in Southeast Asia. Pulse Lab Jakarta (PLJ), a joint data innovation facility by the United Nations and Government of Indonesia, used several population data sources, including HRSL, and advanced analytics to develop indicators to help predict the potential spread of Covid-19, and to prevent the resulting economic fallout. 


  • Using HRSL for Cholera Vaccination of 252,000 children in Mozambique: In 2019, just six weeks after the devastation that followed Cyclone Idai, Northern Mozambique was struck again by Cyclone Kenneth – the strongest tropical cyclone ever to make landfall in the region. Idai sparked a massive Cholera outbreak. While a large vaccination campaign was already underway, there was widespread concern that Kenneth could exacerbate the ongoing epidemic. Prior to Kenneth’s landfall, a multi-sectoral team from Harvard University School of Public Health, Direct Relief, Nethope Crisis Informatics, Facebook Data for Good, and Northwestern University School of Medicine collaborated identify high risk areas and develop a model-based estimation of cholera spread. The goal of the modeling effort was to help prioritize where, given limited time and resources, vaccines should be distributed in Mozambique. The team’s  cholera outbreak risk model used HRSL as an input, which provided population estimates for the areas affected. The team created a web-based tool to help examine these risk factors more closely and predict how affected populations would behave in real time.


  • Universal Access Energy Lab and Waya Energy’s use of HRSL for Rural Electrification Planning: The Universal Access Energy Lab (UAEL) developed the Reference Electrification Model (REM) to help plan detailed electricity networks in developing countries, and connect unelectrified people to the existing infrastructure. REM uses various data streams to determine the optimal electrification modes, estimate costs of electrification for each unit, and produce preliminary designs and recommended engineering systems. As a part of this effort, UAEL uses HRSL to determine the location of residential customers and expected growth through 2024. Waya Energy, a spin-off of UAEL, partners with governments, utilities, agencies, and private-sector organizations to implement this model. Government planners and entrepreneurs can use REM to make better decisions about how to plan and implement electrification projects across the country — including Waya Energy’s work on the National Electrification Plan for Rwanda.  

The Data Artefact Study signals the potential value of population density mapping innovations but also the potential benefits of inter-sectoral partnerships and sharing in the space. The piece also examines several risks and challenges present in the space ranging from poor image quality to the potential for vulnerable populations to be exposed if population information is shared in conflict zones or politically unstable areas. It concludes with an overview of several avenues for future research and experimentation — including last mile service delivery, agriculture forecasting, urban planning, and disaster preparedness.

Read the Data Artefact Study here. 

The Data Artefact Study was informed by stakeholder interviews with Andres Gonzalez Garcia and Clara Pérez-Andújar Carretié (Universal Access Energy Lab); Andrew Schroeder (Direct Relief); Greg Yetman (CIESIN); Andreas Gros (Facebook); and Laura McGorman (Facebook). It also underwent an open peer review process which surfaced invaluable recommendations and insights from Andrew Tatem (University of Southampton); Mrinalini Kochupillai (Technical University Munich); Laura Kahn (Accenture Federal Services); Eduardo Bajar (Fundapi); John McNutt (University of Delaware); Ethar Eitnay (ThinkTank Initiative); and Aleksandar Kostadinov (MCIC).

Back to the Blog

Supported by