Methodological guide on the use of mobile phone data: Dynamic Population Mapping

The Guide on the Use of Mobile Phone Data (MPD) for Dynamic Population Mapping aims to provide practical guidance on how to collect the statistics on population using information from mobile network operators in compliance with United Nations (UN) recommendations on population statistics. It reviews MPD as an alternative (non-traditional) source for providing population data dynamically and filling a gap in measuring the dynamic nature of population presence. This Guide is part of a series of five (5) Guides developed by the UN Committee of Experts on Big Data and Data Science for Official Statistics (UN-CEBD) Task Team on MPD which are designed to guide countries who plan to use MPD for statistics and indicators in displacement and disaster, information society, dynamic population, migration, and tourism.

Preparation of this Guide was led by the organisations Positium and Flowminder with technical support and guidance from the United Nations Statistics Division and the UN-CEBD Task Team on MPD.

Acknowledgement
The UN Committee of Experts on Big Data and Data Science for Official Statistics has prepared guiding materials for the use of mobile positioning data for dynamic population mapping for the statistical community. The Guide on the use of Mobile Phone Data (MPD) for Dynamic Population Mapping has been completed through a consultative process of experts who have years of experience working with MPD.

The main authors from Positium, Flowminder, University of Tokyo, BPS Statistics Indonesia and the World Bank have a combined 50 years of experience with MPD with them. The Guide reflects the authors' knowledge and ideas for measuring population. In many cases, it reflects the experience of whole organizations with this dataset, making this Guide one of the best researched sources on the topic. Others, mainly from UN organizations have applied their experience working in the domain of population statistics and have provided comments that have been instrumental in taking the ideas to an organized final version.

Table of contents

• Abbreviations
• 1. Introduction
• 2. Background and context
  • 2.1. Traditional versus new data sources
• 3. MPD for dynamic population mapping - key considerations
  • 3.1. Operational aspects
  • 3.2. Privacy and ethical aspects
  • 3.3. Quality aspects
  • 3.4. MPD limitations
• 4. Data types and forms
  • 4.1. Data types
  • 4.2. Data forms
• 5. Applications and uses
  • 5.1. Mobile positioning data for studying population distributions and movements
  • 5.2. Resident (de jure) population mapping
  • 5.3. De facto population mapping
  • 5.4. Monitoring population redistribution as a result of COVID-19 mobility restrictions
  • 5.5. ‘Dynamic denominators’ to estimate disease incidence
  • 5.6. Disaster preparedness planning and response
  • 5.7. Infrastructure and resource planning
  • 5.8. Creating dynamic sample frames for surveys
  • 5.9. Census
    • 5.9.1. Census date and interview time preparation
    • 5.9.2. Census enumeration area preparation
    • 5.9.3. Undercounting mobile populations
    • 5.9.4. Daytime population
    • 5.9.5. Intercensal population estimates
• 6. Methodological approaches and challenges
  • 6.1. Data modelling approaches
  • 6.2. Components of dynamic population mapping
    • 6.2.1. Detecting place of residence - home detection algorithms
      • 6.2.1.1. Simple tower-level approaches
      • 6.2.1.2. Clustering approaches
      • 6.2.1.3. Anchor point models
    • 6.2.2. De facto population concepts
    • 6.2.3. Stay detection methods
    • 6.2.4. Inferring population estimates from subscriber counts
      • 6.2.4.1. Number of SIMs versus number of subscribers
      • 6.2.4.2. Representativity
      • 6.2.4.3. Modelling population estimates
• 7. Quality Assurance
  • 7.1. Assuring quality of input data
  • 7.2. Assuring quality of the statistical process
  • 7.3. Validating methods and assuring quality of output
• 8. Case studies
  • 8.1. Estonia: De facto population statistics for Estonian Police and Border Guard 2012
  • 8.2. Namibia, Dynamic Denominators_ the impact of seasonally varying population numbers on malaria incidence estimates
  • 8.3. Guinea, Liberia, Sierra Leone: Dynamic population mapping including cross-border movements
• References