Using Mobile Network Operator Data for COVID-19 Response
14 mins read
DIAL has developed this introductory brief as the first in a series of materials on using mobile network operator (MNO) data.
This page was last updated on June 17, 2020.
MNO Data | MNO Data for COVID-19 | Getting Started | Key Terms | Contacts and Resources
DIAL has developed this introductory brief as the first in a series of materials on using mobile network operator (MNO) data. When combined with other traditional datasets, MNO data can be analyzed and used to support development policymaking – including emergency response efforts around COVID-19.
Our aim is to provide governments and intermediaries with openly licensed, editable resources on:
(1) The opportunity of mobile network operator data;
(2) How to get started with using mobile data;
(3) The various perspectives of the groups stakeholders involved; and
(4) How to bridge the demand side (NGOs, governments, and multilateral organizations) and the supply side (MNOs and aggregators) of this work.
These initial resources are tailored towards a country’s COVID-19 response, though they can also be customized and used to support other long-term public health and development goals. This introductory brief provides an overview of using mobile network operator data as part of a country’s COVID-19 response. We hope it serves as a useful primer for governments and intermediaries.
Mobile Network Operator Data
Mobile phones generate enormous amounts of data every day as people move around and connect to various networks. This data, which is collected by mobile network operators (MNOs), can help fill the gaps and improve our understanding of population movements over time. When MNO data is combined with other types of data, such as satellite data and health data compiled by the government, we can produce valuable insights to help us deal with a range of development and humanitarian issues, including the COVID-19 pandemic. Many governments and intermediary agencies are already using MNO data in order to improve health outcomes, bolster their disaster preparedness and prevent food shortages.
In low-income countries, where national statistics may be under-resourced and data quality poor, MNO data analytics can help us understand mobility patterns. MNO data is anonymized and aggregated to generate national and sub-national patterns and trends about population movement. These near real-time insights into mobility patterns can provide critical information to enhance decision-making about health services and lockdown measures and help model the impact of COVID-19.
The accuracy of the mobility patterns will vary depending on the number of cell phone towers and the mobile penetration rate. Some areas, often rural areas where the poorest and most vulnerable live, have lower connectivity rates than others, so individuals living in these areas may be under-represented in mobile data. In order to mitigate and account for these limitations, MNO data can be combined with other types of data and verified against other data sources when possible.
Data privacy and security are an essential part of using MNO data, and DIAL works with governments and intermediaries to ensure that the benefits of using MNO data outweigh the potential privacy and security risk. We work alongside them to promote safe and responsible data use to enhance their ability to deliver public services to those who need it most. As a starting point, where MNO data is being used, DIAL recommends that personal data never be shared in its raw form and any data containing personal data should be anonymized and aggregated before processing. Embedding strong privacy and security standards and practices into the project provides safeguards so that MNO data can be used responsibly.
MNO Data for COVID-19
When combined with other datasets, MNO data analytics can help governments understand short-term, long-term and seasonal mobility patterns, which can help decision-makers to plan health interventions at various stages during the pandemic.
Mobility patterns that are generated from MNO data can help decision-makers to understand:
- Virus spread. Mobility patterns can be used as inputs to model the risk of contamination of the population throughout the country.
- Hotspots. Mobility patterns can help identify key transmission or transit nodes, as well as particularly at-risk areas. For example, several European governments are using MNO data to predict potential hotspots of COVID-19 infection.
- Health services. When mobility patterns are combined with health data, they can help Identify health facilities that are at risk of being overwhelmed by patients and understand how to distribute medical resources, such as testing kits, beds, and medical staff. Cooper/Smith is working on using MNO analytics to understand health system capacity during COVID-19, predict where new infections are most likely to crop up, and see where healthcare providers will likely be strained and need support.
- Social distancing. Mobility patterns can help governments monitor whether there were changes in mobility patterns following the announcement and implementation of social distancing and other control measures. Flowminder points out that this type of monitoring can be implemented at a relatively low cost, making MNO data a particularly useful tool for countries with limited resources.
- Economics. We can measure short-term and long-term changes in mobility patterns following the onset of COVID-19. These mobility patterns can be used to assess the short- and long-term social and economic consequences of movement restriction measures and disease migration.
- Other uses? New uses are emerging as research and evidence develop. If you are working on new uses of MNO data analytics, we would like to hear from you. Contact us!
Last year, DIAL partnered with Cooper/Smith, the Malawi Ministry of Health, and MNOs in order to improve access to healthcare in Malawi. Below is an example of how MNO data was used by the Government of Malawi.
Example: MNO data and health clinic placement in Malawi
Malawi’s population is growing rapidly, and half of its people do not have access to health services. Additionally, many people migrate during the rainy season, making it challenging to understand where people are located over time. As a result, the government of Malawi decided to use MNO data to better understand the population movement and improve access to health services for all citizens. MNO data was used to map population and migration patterns, which were then compared with other types of information in order to produce insights. (See Figures 1 and 2). From these insights, the government was able to figure out the best places to build new health clinics across the country so as to reach the maximum number of people.
Source: Our Data Work In Malawi
Considerations for Getting Started
First, you need to clearly define the problem you are trying to solve and identify how the insights from MNO data will be used. It is critical to understand the logic flow that links the MNO data and analysis to the decision-making and ultimate impact that is intended. Consider how you will use the analytics towards an effective COVID-19 response. Consider how you will link the data insights and the analysis to the decision-making and the ultimate impact that is intended. Is the intended user of the data analysis – for example, the MoH – able to absorb and act on this new information?
It is also vital to balance any potential benefits of using MNO data against potential risks or harms that may negatively impact citizens – both now and in the future. The Principles for Digital Development can help you design an effective and ethical intervention using technology.
DIAL works to ensure that all stakeholders have access to best practices and the tools to use MNO data in a safe and responsible way and generate insights that can enhance the delivery of services to the people that need them most. Below are five considerations for getting started.
(1) Is there political support for using MNO data?
Identify if the government is interested in using MNO data as part of its COVID-19 response. Reach out to the ministry of health to assess its interest, as it will be essential to get its approval to perform this work.
There are various other stakeholders involved in this kind of work, and their cooperation and support are vital to the project’s success. In order to gain their support, consider each stakeholder’s role, responsibilities, incentives and constraints in your data for development (D4D) project.
To get initial buy-in from stakeholders, consider hosting a virtual meeting to discuss and demonstrate the potential uses of MNO data for the COVID-19 response. The meeting should include the ministry of health, government officials, the telecommunications regulator, the national statistical office, MNOs, and a technical partner. Consider also including civil society organizations, international organizations and NGOs, and legal experts. Following the meeting, assess the political will of the stakeholder group to support and engage with this work.
Identify existing digital platforms and systems that are using mobile and/or health data in your country. Is someone at another government department, international organization, private-sector company or NGO already working on this in your country? If so, reach out to them to collaborate and find partners, as they will have likely already completed much of the preliminary work needed to use MNO data as part of the COVID-19 response.
Identify Potential Partners
Intermediary organizations play a facilitating role by convening project stakeholders together. Intermediary organizations include Data Pop Alliance, DIAL, Flowminder, GSMA, and UNICEF. Contact DIAL (see contact information below) and we can help put you in touch.
Technical organizations help identify, analyze, and process data to generate analytics and insights. Technical partners include Cooper/Smith, Dalberg Data Insights, Data Pop Alliance, Flowminder, and the University of Tokyo. Technical capacity may also exist within an MNO or aggregator, or the telecom regulator itself.
Implementing organizations helping to plan and manage the project and carry out the day-to-day program activities on the ground in-country. These tend to be NGOs and include Cooper/Smith, PATH, UNICEF, and Village Reach.
(2) Is there mobile and health data available?
When trying to locate and access relevant datasets, ask what are you trying to find out, and why is this important? What information might help you answer this question? Consider which datasets already exist, where are they stored, and what would be needed to combine with the MNO data to create useful analysis.
MNO data holders include MNOs and aggregators. Health data can be found through the Ministry of Health and NGOs operating in country at the national, regional or local level. Country-level data can be found at the National Statistical Office and within relevant government ministries. There are also publicly available datasets, such as World Health Organization Workforce Statistics and OpenLMIS that can also be used and combined, if available.
In order to produce useful insights for COVID-19, MNO data needs to be combined with other datasets, such as health and census data. These data sets must be legally accessible, of sufficient quality, and available to use. Consider the following:
- Can you legally use MNO data? Assess the privacy and technology deployment protocols and laws and reach out to key government points of contact in order to understand which laws cover the use of mobile data. Identify which regulations set limits and provide guidance on the use of MNO data. Plan out what steps are required to get approval from the health ministry or telecommunication regulator to get access to MNO and health data to carry out this work. Reach out and begin to liaise with the legal team in the MNO, to start consultations on requirements and constraints for using MNO data.
- Do you have sufficient mobile connectivity? Assess the mobile connectivity in critical regions of the country and identify which and how many mobile operators would have to participate to sufficiently cover the target population. Information on the mobile penetration rate is often readily accessible from the telecom regulator.
- Can you get mobile operators to share their MNO data? Identify how to incentivize MNOs to share their data. For example, MNOs are often willing to donate their data for free in exchange for an opportunity to build their own internal capacity and infrastructure or generate insights for their internal business intelligence. Reach out to the dominant MNO market leader, as they will be able to provide data for a significant portion of the market. However, keep in mind that data from additional MNOs will provide a more accurate understanding of population density. Assess the MNO’s internal technical capabilities in terms of data management, analytics, and the ability to anonymize data prior to sharing it with a government or technical partner.
It is important to keep in mind that this work can be resource intensive for MNOs. At the moment, they are willing to divert resources within the MNO to help with the COVID-19 response. However, this will be challenging for MNOs to sustain over the long term. COVID-19 disease suppression and economic recovery is likely to require careful management over a period of many months or even years. Sustainable long-term funding arrangements are advisable – as this funding will enable MNOs to continue to provide relevant data and analytics over the longer-term duration of the pandemic. These analytics can then help inform government policy decisions on interventions that affect people movement in a dynamic way; and enable the effectiveness of policies to be measured over the full COVID-19 recovery period.
- Do you have quality health data available? In addition to MNO data, you also need to identify, locate and evaluate health-related datasets. First, identify which specific health datasets are useful and necessary to answer your questions on COVID-19, such as data on health clinic distribution, the health supply chain, immunization distribution and disease burden. Do a quick scan of the health data quality and accessibility. Is the health data integrated into one data system accessible by the ministry of health (e.g., on a health management information system such as DHIS2)?
- Is there additional data available? Identify whether there are other datasets available that can supplement the existing health and MNO data, such as census, satellite imagery, and other geospatial demographic datasets like WorldPop and GRID.
(3) Are there data sharing agreements in place?
When engaging in MD4D work, the sharing and use of MNO data must be rooted in mutual understanding, trust, and obligation. This requires clear understanding from government stakeholders, technical service providers, and implementing partners about data governance, privacy and security at each step of the data life cycle, including transfer, storage, processing, and analysis. Partners can strengthen trust, transparency and accountability around accessing, sharing and using data through the use of Data / Analytics Sharing Agreements and Data Processing Agreements.
Data sharing agreements establish a framework and mechanism to share data between the project partners. Agreements can take various forms and might address data sharing or data processing arrangements. Agreements must comply with local laws and regulations and should clearly and accurately set out details regarding data management arrangements including transfer, ownership, storage, processing roles and responsibilities, and security and privacy cooperation. Agreements should outline specific objectives for the project, including details on how the data will be used to support the Ministry of Health or other government agencies in their COVID-19 response. Agreements should also address other key issues including, intellectual property, choice of law, and dispute resolution.
Contracts for Data Collaboration has a repository of template agreements for practitioners to reference, customize, and use, as well as an analytical framework to help parties better understanding of the nature and function of data sharing agreements.
(4) Is there public support for use of MNO data?
Transparency and communication about using MNO data will be vital. It is essential to have a strong foundation of public trust in government data in order for MNO data use to be an effective tool for a COVID-19 response. The government and public need to understand what kind of data is being shared, how it is aggregated and analyzed, how de-anonymization is prevented, and how the outputs and insights will be used. Clear accountability and impact monitoring are key. Investing in explaining the technology to both subscribers and government agencies that will rely on these insights builds comfort and buy-in.
At the outset, consider consulting or setting up an appropriate ethical review body or engaging a partner to support an Internal Review Board process. These bodies can help assess whether there are ethical, civil liberties or human rights issues that will come into play, either now or in the future.
** Note: It is important to clarify that this resource covers the use of aggregated MNO data to understand trends in population dynamics. DIAL does not advocate for the use of MNO data to identify individual exposure risk based on proximity to known or suspected cases, also known as contact tracing. Using these approaches for this purpose would produce inaccurate and misleading findings.
Aggregated data – Summarized data representative of a larger dataset or population. Aggregated data does not contain personally identifiable information (PII).
Analytics – Analysis of current and historical data that is presented using charts, graphs or maps. Algorithms are often used to process data and generate analytics. (OCHA)
Anonymized data – Data that has irreversibly removed or modified personally identifiable information, so that a person can no longer be identified directly or indirectly. (OCHA)
Call data records (CDRs) – A record of a voice call, text message or other transaction. Call data records include the mobile numbers of those making and receiving the call, date, time, call duration, and low-resolution location information (nearest cell tower). (GSMA)
Insights – The product, dashboard or visualization that is generated from reading analytics. (GSMA)
Mobile data for development (MD4D) – An analysis of CDRs Call Detail Records (CDRs) that provides valuable information for development action. (UNGP)
Mobile network operator (MNO) data – Data generated from mobile network operators, such as call detail records, profile data (age, gender, etc.), location data (number of people in specific locations), usage (number and duration of calls and text messages) and spend data (monthly charges, currency, on-time payments).
Personal data – Any information relating to an identified or identifiable person (data subject). An identifiable natural person is one who can be identified, directly or indirectly, particularly by reference to an identifier such as a name, an identification number, location data, an online identifier, or one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that person. (GDPR)
Personally identifiable information (PII) – Also called direct identifiers. This type of information (e.g., name, social identity number) reveals directly and unambiguously the identity of a data subject. (OCHA)
Pseudo-anonymization – Does not remove all identifying information from the data but reduces the linkability of a dataset with the original identity of an individual (e.g., via encryption). (GDPR)
Re-identification risk – The risk that an individual can be identified by combining an anonymized dataset with other types of information. (OCHA)
Contacts and Resources
Convene the right team. Governments can work with intermediaries like DIAL to bring key stakeholders together, support technical work and legal arrangements, and ask the right questions. Many potential partners developed the resources that we have linked to on this page.
If have questions about how to get started or if you would like assistance getting connected to the right partners, we would be glad to help:
Rachel Sibande, firstname.lastname@example.org
Tanvir Singh Natt, email@example.com
Claude Migisha, firstname.lastname@example.org
Diana Sang, email@example.com
Contracts for Data Collaboration
Data Sharing Agreements – Detailed look at how to create data sharing agreements
Data Sharing Agreements – Repository of example agreements and analyses
How to use your data to fight COVID-19 – Roadmap for how to implement a D4D COVID-19 response in Africa
Data Pop Alliance
Global South Response and Recovery – Strategic Vision and Action Plan
Resources for COVID-19 response – Overview of tools for COVID-19 response
MNO and Aggregator Catalog – Repository of MNOs by country
Pooling Demand for MNO Data – Overview of supply and demand of MNO data
Unlocking MNO Data – Overview of the uses and challenges of using MNO data
Using MNO data for COVID-19 – Knowledge center with code, methods, guidelines and overview of uses of MNO data for COVID-19
Mobility analysis to support the Government of Ghana in responding to the COVID-19 outbreak
COVID-19 Privacy Guidance – Recommendations on how MNOs can maintain trust while responding to requests for insights from MNO data
Policy and Regulatory Best Practices for COVID-19 – Forum to submit, share and read about policy and regulatory best practices related to COVID-19
WHO – Digital Health Atlas – Repository of global technology platforms and systems
COVID-19 Digital Preparedness – Useful advice on creating local asset logs and context analyses in the early days of response
COVID-19 Resources from National Statistical Offices – Repository of best practices and resources
This D4D resource for COVID-19 response by Digital Impact Alliance (DIAL) is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.