Below we include some excerpts from the guide, including the introduction. A short Q+A on the guide is also available. Download the complete guide for all the details. 

Introduction

Understanding how, when, and where people move is especially important during a public health crisis. In many countries in the Global South, there is a lack of robust transit or mobility data to help ground-truth mobility patterns. Even in the cases where these data exist, they usually contain one-time snapshots rather than time series views of mobility.

That’s why call detail records (CDRs) have become a popular source of mobile phone data for researchers and policymakers. CDRs are logs of network events (voice call or SMS) that are made on the network of a telecommunications provider. Each CDR includes the cell tower the subscriber was closest to, so the locations of cell towers can be used to approximate population mobility patterns. Importantly, CDRs are automatically generated by mobile network operators, so they do not require additional infrastructure or effort to collect like other types of mobile phone data do.

Research using CDRs ranges from mapping population movement to understanding migration patterns, exploring community structure, and understanding disease spread. 2 These kinds of studies demonstrate that CDR data can be used to understand population mobility patterns, the temporal dynamics of mobility, and the relationships between socioeconomic status and mobility.

This guide provides an introduction of how to use CDRs to inform public policy response during a crisis. It draws on our research team’s experience working with CDRs in Sierra Leone during the early stages of the Covid-19 pandemic. We have distilled our process into three steps that any team interested in using CDRs should consider:

Step 1. Properly accessing and securing data. Step 1 entails getting access to the CDR dataset and storing it in a secure environment. Another important task here is ensuring that no sensitive or personally identifiable information is released —such as a name, address, or phone number.

Step 2. Checking the data quality. Step 2 focuses on ensuring that the records are valid, accurate, complete, and consistent.

Step 3. Facilitating analysis and communicating results. Step 3 is where the analytics team chooses the types of analysis they will run and configures an environment to support the processing of the data and the communication of the results.

The goal of this guide is to provide a detailed overview for nontechnical policymakers who want to learn more about how CDR analysis works and what to expect throughout the process. It covers technical details of the analysis, with the aim of helping all partners collaborate effectively and understand their roles. The last section of the guide discusses data ethics, privacy, and the limitations of CDR data.

Background: Mobility Study in Sierra Leone

Sierra Leone reported its first case of Covid-19 on March 30, 2020. Soon after, the Government of Sierra Leone (GoSL) implemented a three-day lockdown, followed by a 14-day, interdistrict travel ban. As the pandemic developed, the government wanted to understand the effectiveness of their mobility restrictions and subsequent policy actions—did the lockdown measures actually change people’s mobility patterns?

To find out, a research collaboration was formed between GoSL’s Directorate of Science, Technology and Innovation (DSTI), Africell (the mobile network operator with largest market share), and researchers at the Massachusetts Institute of Technology’s Civic Data Design Lab and MIT Governance Lab (MIT GOV/LAB) to answer the following questions:

  • Did the three-day lockdown and the travel ban decrease mobility in Sierra Leone?
  • If there was a change, what were the differences across districts?

The researchers tackled these questions using call detail records (CDRs), logs of cell tower pings that can serve as a proxy for mobility. A summary of those findings with interactive data visualizations is available online, while a book chapter in “Urban Informatics and Future Cities” covers the research in more detail.

Suggested Citation: Ndubuisi-Obi Jr., Innocent. 2021. “Making the Most of Mobility  Data (CDRs): A Guide for Policymakers”, MIT Governance Lab and Civic Data Design Lab (United States).

Thanks to our partners at Sierra Leone’s Directorate of Science, Technology & Innovation (DSTI), Africell-Sierra Leone, and Flowminder Foundation, including Aurelie Jeandron, Veronique Lefebvre, and  Xavier Vollenweider. And, thanks to Sarah Williams, Lily L. Tsai, Alisa Zomer, Will Sullivan, and Maggie Biroscak for providing input into various versions of this guide. Illustrations and graphic design by Susy Tort and Gabriela Reygadas.