Building capacity to generate statistics from mobile phone records to support COVID-19 response

5 min read
Share on facebook
Share on twitter
Share on linkedin
Share on email

A few weeks ago, UN Global Pulse organized a series of webinars meant to assist National Statistics Bureaus in Africa with using mobile phone data to respond to, and to support recovery from, the COVID-19 pandemic. Although focused on the African context, the learnings from these webinars could be applied by statisticians in national offices across geographies. 

Although the generation of official statistics from big data sources remains limited, National Bureau of Statistics (NBS), which are the custodians and main producers of data at national level, are increasingly experimenting with the use of big data, especially in the context of the COVID-19 pandemic. In response to requests from statisticians in Africa, UN Global Pulse conducted a series of webinars to build the capacity of NBSs to use anonymised mobile phone data for decision-making in response to this pandemic and future crises. 

Throughout the course of 10 sessions, experts from the private sector, academia, and the UN discussed applications of mobile data with practical implications for NBSs. The sessions were interactive, combining theory with examples of data-driven projects, hands-on coding, and guidance around data privacy, protection and ethics. 

A total of 85 participants joined the seminars from 13 NBSs across Africa and not only:  Nigeria, Uganda, Rwanda, Senegal, Kenya, Somalia, Ghana, Indonesia, USA, Philippines, Ethiopia, Zimbabwe, United Kingdom, Nepal, Georgia, Gambia and Burundi. 

Building capacity to generate statistics from mobile phone records to support COVID-19 response 1
GSM call flow and CDR generation/Telecom data

What participants had to say

“…My capacity and confidence has been built, to be able to analyze anonymized mobile phone data for COVID-19 and other responses….”. 

CDR data holds much more value than anticipated. Although the previous is true, there is still bias and unknowns within this data format which needs to be considered. 

The use of mobile phone data for COVID-19 response shows that data generated by telecommunications companies are not limited to the use of the business, but can also be extended and used by the government to provide better services and policies.”

Public and private entities should go hand in hand in times of crisis. We can take this step further to handle the pandemic better.

An overview of the webinar series 

Module 1 – Diving into the topic. The module introduced the concepts of data science, AI, machine learning and big data for the analysis of call detail records (CDRs). It included an in-depth presentation of mobile phone data sets (composition and registries). Participants learned how to perform statistical interference on mobile data to extract features and generate statistics as daytime population. Tools for processing CDRs were presented and participants were challenged to use Python. 

Module 2 – Engaging in an in-depth analysis. The module presented the components of the mobile phone data processing pipeline, including identification and extraction, data exploration, data cleaning, data anonymisation and analytics. Interactive sessions were hosted to present the guiding principles for using visualizations and to guide the definition of statistical indicators from CDRs. Practical sessions were held to familiarize participants with the Hadoop ecosystem and other tools to analyse mobile phone data.  

Module 3 – Framing analysis results. The module examined data privacy and ethical risks associated with the use of CDRs. An interactive session presented biases and limitations of mobile phone data to generate statistics, including the demographic bias, limitations related to accuracy in space and time, and the effects of behaviour. The module continued with the presentation of an in-depth analysis on how to generate statistics to monitor the socio-economic impact of COVID-19 with a subset of mobile phone data, namely mobile money transactions. Practical sessions were held to help participants learn how to code with mobile phone synthetic data. 

See the full agenda. 

How mobile phone data can inform COVID-19 response 

The COVID-19 pandemic has generated an unprecedented chain reaction in individual and societal behaviour around the world. These changes were driven by an array of factors including reduced human mobility, limited social interaction, adaptability to teleworking or home schooling. 

In a digitalized world, changes in the behaviour of individuals and communities leave a digital print. The COVID-19 pandemic has spurred a digital boom, which is manifesting itself in the use of tablets in hospitals to help patients feel closer to family members, in children attending on-line classes , or in the surge of e-commerce for basic goods. Before this boom, the exponential growth in access to mobile phones was already narrowing the digital divide. The number of mobile-cellular telephone subscriptions are now greater than the global population and 96 per cent of the world’s population lives within reach of a mobile cellular network (International Telecommunication Union). The pandemic exacerbated the use of mobile apps across age groups, gender, ethnicity and geographical areas. Along with the digital boom, the number of digital prints exposing changes in societal behaviour also multiplied. Analysing the effects of these changes at scale is critical to monitoring and measuring the socio-economic impact of COVID-19.

Mobile phone data is one of the largest treasure troves of big data, and different actors are tapping into the granularity of information derived from anonymized call detail records in the fight against the pandemic. For example, it can help identify regions and population groups where vulnerabilities are intensified, detect changes in livelihoods, and variations in regional dynamics. It can help monitor coping strategies and the levels of resilience of vulnerable groups to inform relief and recovery efforts once the health emergency has passed.

Links

UN Global Pulse designed the webinars building on knowledge acquired during extensive work with anonymized CDRs.  

———

UN Global Pulse would like to thank the German Corporation for International Cooperation GmbH (GIZ) and the William and Flora Hewlett Foundation for their continued support to this activity and UN Global Pulse’s overall mandate.

Did you enjoy this blog post? Share it with your networks!

Share on facebook
Share on twitter
Share on linkedin
Share on email

News, thoughts and ideas about big data and AI, data privacy and ethics from across the Pulse Lab Network. Read more on the blog.

Pulse Lab New York

UN Hosts First AIS Big Data Hackathon

On 3-6 September 2020 the United Nations Statistics Division, along with UNCTAD, UN Global Pulse, Marine Traffic, and CCRi, hosted the UN’s first AIS Big

Scroll to Top