Donors, development agencies and other international organisations are continuously investigating new opportunities to employ digital technologies to improve the ways they design, implement, and measure progress with development projects. During the Mobile 360 Africa Conference in Dar es Salaam, the GSMA hosted a panel discussion to discuss this topic in greater detail with Nicolas de Cordes from Orange Group, Eric Anderson from the World Bank, and John Quinn from UN Global Pulse.
The panellists on this ‘Data for Development’ session highlighted several actual and emerging use cases, as well as key challenges when mobile-derived Big Data are used to achieve public policy objectives and to support projects that impact the poorest segments of society. In particular, our panellists outlined how mobile operators’ Call Detail Records (CDRs) could be used to provide development stakeholders and policymakers with meaningful information in real-time, leading to better interventions and decision-making.
Emerging Use Cases for Mobile Data
Due to the prevalence of mobile phones in developing countries, CDRs could provide institutions with up-to-date information on large swathes of the population, including many households that are found at the bottom of the pyramid. According to the Data-Pop Alliance, large mobile operators collect over six billion CDRs per day, and each record contains useful information on individual mobile subscribers including the time, location and duration of their calls. When appropriately aggregated and anonymised, this rich data could provide valuable insights that help government ministries and development organisations achieve important public policy and development objectives – from tracking the spread of diseases to addressing high levels of pollution.
Our panellists agreed that the development sector is increasingly aware of opportunities to leverage CDR data for development purposes, but suggested that most organisations are still in an exploratory, transitional phase. With that in mind, they highlighted a few ways CDR data could impact development projects in the near future.
There was a lot of excitement around opportunities to leverage caller mobility data, and using this to augment other available information and develop better, more reliable services. For instance, data on commuter movement – such as when and where commuters begin and end their journeys – could be used to improve public transportation services and urban planning. Nicolas de Cordes pointed to French tech company SNIPS, which has already developed contextual models that predict the risk of car accidents in major cities by looking at features such as traffic, street topologies, proximity of bars, the road surface, or the weather.
Mobility data could also improve access to clean water. Edward Anderson referenced an estimate that 50 per cent of water points in Tanzania are non-functional. By analysing mobile users’ movement patterns to and from various water points, government ministries could gain a much clearer picture of where breakages are occurring; a sudden drop in traffic around one water point, paired with increased traffic to neighbouring water points, would likely indicate a malfunction.
Lastly, panellists suggested that CDR data could provide up-to-date statistics on population density and traffic movements, assisting ministries as they plan infrastructure projects to provide better access to new roads, water, electricity, and health services. For instance, a health ministry investigating where to build a new hospital might look for highly populated areas that are more than thirty minutes away from existing emergency services.
Remaining Challenges, Next Steps
The crucial challenge raised by each of our panelists is the asymmetry between the supply of, and demand for, Big Data. As John Quinn pointed out, ‘just because the data is available doesn’t mean everyone is ready to use it’. Experience from the World Bank and UN Global Pulse has shown that it is still challenging to find the right users for CDR data, such as urban planners or task forces that are fighting disease. Government ministries often lack the capacity to perform a robust analysis of CDR data, and it could be quite some time before government-led data analysis evolves into proper ‘data science’. There is also some uncertainty around whether front-line workers (such as development practitioners or health workers) are ready or willing to change the way they operate as a result of data analysis.
Similarly, from Orange’s perspective there is still not a substantial market for CDR data because NGOs and government ministries are still unsure about how or why to ask for the data, while operators are considering how aggregated the data should be to assuage concerns around privacy. Vital questions remain about who the data belongs to, and how institutions can ensure that the privacy of individuals is protected. Mobile network operators (MNOs) are well-placed to understand the potential risks to individuals and groups from Big Data analytics and can implement measures to avoid or mitigate those risks. Even so, negotiations between MNOs and other institutions over the provision and use of data can be both difficult and lengthy.
To address these barriers, the Global Partnership for Sustainable Development Data and the Data-Pop Alliance are helping to facilitate policy debates, partnerships, and grassroots advocacy efforts to promote the use of Big Data in development. Orange Group has hosted a ‘Data for Development’ innovation challenge in Senegal to help advance the use and analysis of CDR data to create social impact. These organisations are now teaming with the Massachusetts Institute of Technology and World Economic Forum through the OPAL project, which aims to engage data ecosystem (providers, users and analysts) to build local capacities and connections and foster innovation.
The GSMA is also making significant progress in this space. In the wake of the Ebola outbreak in 2014, we worked with the mobile industry to develop guidelines on the protection of privacy in the use of mobile data which outlines, in broad terms, the privacy standards that mobile operators must apply when subscriber data is used in exceptional circumstances. We are now working to revise the guidelines to make them more broadly applicable to Data for Development use-cases, such as anticipating and tracking the spread of dengue fever in Pakistan.
The GSMA has recently developed a policy position on the use of personal data for social impact, advocating for the use of big data to create social value while still protect a person’s digital identity. Some of the steps mobile operators can take to address privacy challenges and increase trust between industry stakeholders and consumers include:
- Finding innovative ways to provide meaningful choice, control and transparency to individuals about what data is collected and how it is used.
- Thinking carefully about the impact of the insights derived from Big Data on individuals (and groups), and the actions or decisions that may be taken based on those insights.
- Establishing clarity on responsibilities between parties when collaborating on Big Data analytics projects.
Over the next several months we will be working with the UN Global Pulse to identify where and how the Data for Development community has evolved, what some of the meta barriers have been and what we expect to see in the future. By working closely with development partners we believe the mobile industry will help facilitate a shift in the sector’s ability to leverage big data and address some of the world’s most pressing social needs.