The intention of this exploratory module is to explore the possibilities afforded by big data for systematic research on development-related questions. The agreement of mobile operators (and others who have access to other streams of big data) is essential for the conduct of the research. The intention is to conduct a pilot project that can be used to obtain the consent of big data owners in the future. Another related activity will be that of developing guidelines for the ethical use of mobile network big data in collaboration with mobile operators. The development of the guidelines creates a natural platform for a conversation with operators on terms of access to the data. This activity will also contribute to the overall objective of creating the conditions for greater access to big data for development research purposes.
The proposed research seeks to identify communities based on actual communication patterns rather than administrative boundaries. Sri Lanka, according to official figures, is not a highly urbanized country, with only 15-21 percent of the population living in urban areas.
Urbanization and growth go together: no country has ever reached middle-income status without a significant population shift into cities. Urbanization is necessary to sustain (though not necessarily drive) growth in developing countries and it yields other benefits as well (Annez & Buckley, 2009).
There are 23 municipalities, including Colombo, the largest with a population of around 800,000. All are unable to provide citizen services at the required levels, primarily because of poor administration and low revenues. Municipal demarcations have been inherited from colonial times and have little relation to the actual functionings of the cities at present. The objective of the proposed research is to contribute to an evidence-based discussion on the optimum demarcations of cities and related public-policy issues, using findings from the analysis of big data from two major mobile operators who have consented to cooperate.
Examples of the policy-relevant questions that we hope to address are:
- How integrated are the different parts of cities?
- Is the integration between parts of cities and suburbs greater than between parts of cities?
- How far do those who work in cities commute?
- What are the optimal locations for delivering government services?
A core set of research questions that can be answered using only mobile data will be developed. Given the conclusion of the Sri Lanka census in 2012, best efforts will be made to combine mobile data with census data to obtain richer results.
The data will be analyzed using T-Cube, a software package from Carnegie Mellon University that has been used for pattern analysis in previous LIRNEasia research. In addition, software packages currently in use by the mobile operators may be used subject to negotiated conditions.
Development of a self-regulatory regime
In parallel with the work on using mobile operators’ big data to answer social science questions, LIRNEasia proposes to develop a self-regulatory regime for the collection and use of mobile data. This would be consistent with the respective country’s laws and regulatory provisions, but would not be imposed by government. Given the multinational nature of mobile operators and the inherent mobility of data, there is value in striving for inter-country consistency.
This activity will draw on basic principles of good practices related to privacy and on the practicalities of working with big data. Here, the project will draw upon the output from, and resources within, the privacy related research networks being developed under IDRC auspices. Draft codes of practice will be developed for discussion with the direct stakeholders. This activity will serve two purposes: incorporation of the knowledge of the operators and the education of relevant officials on the need to safeguard privacy principles. Given the sensitivities of the subject no formal commitment will be made with regard to the establishment of an implementing mechanism, but best efforts will be made to ensure that such a mechanism is set in place.