← Go back

Data | Algorithms | Policy

The LIRNEasia DAP team is the continued form of the Big Data team. Established in 2013, the Big Data team pioneered in the use of large datasets for policy in the Global South – particularly using call data records to great effect. As times change, however, and as computing advances, we found ourselves engaging with increasingly sophisticated algorithms and their effects just as much as with the underlying data they operated on, and expanding beyond the analysis of data into questions of algorithmic bias, the ethics of machine learning, and ever-more nuanced economic and social science into our work. The name Data, Algorithms and Policy thus reflects the implicit scope of our work as well as our end goal: to channel the fruits of computer science into better policy outcomes.

Privacy is still a key issue when it comes to data for development. Often, comparison across multiple datasets can render any measures taken to mask or remove identification of individuals, redundant. In one particular case of our ongoing partnership with the University of Tokyo, though, Viren and Lasantha dabbled in some data generation to overcome the issue. It turned out to be harder than we expected but we kept trying. In July 2019, we submitted a paper on the findings which was accepted as a poster, to NetMob in Oxford.

In July 2018, Lasantha and Viren also started looking at internal migration. Initially, the exploration was focused on ego networks (the connectedness of an individual), based on the hypothesis that the more connected an individual, the more likely they are to engage in positive migration and vice versa. This later evolved to include other characteristics available in the dataset. We derived some parameters from other work on social and spatial diversity and entropy and are developing our own entropy-based parameters, in order to predict internal migration.

Scope: Entire Country

A bivariate choropleth map of internal migration in Sri Lanka. Inward and outward migrations are represented by pink and green univariate color scales respectively. Both color scales have been rendered on the same map and the blended color—purple—represents both inward and outward migrations. Click or tap on a map-segment to change the scope, which then recolors the map to show the inward and outward migrations with respect to that map-segment. Double-click or double-tap anywhere on the map to reset the scope to the entire country.

As a by-product of the migration work, we also developed a socio-economic index for Sri Lanka. This was based on the national census of 2011, and replicated work done in the UK, Europe, Australia, New Zealand, and the UNDP 2019 Global Multidimensional Poverty Index, among others.


A choropleth map of the socioeconomic index (SEI) for Sri Lanka. Higher SEI values represent better-off areas and lower SEI values represent worse-off areas.

Privacy is one part of a larger global discussion on ethics related to data, algorithms, and artificial intelligence. And like much of the conversations on ICTs and digital life, the conversation is skewed towards the global north. In July 2019, Ramathi began studying debates around the ethics of data, AI, and algorithmic decision making, specifically how these issues may be different in the global south vs. the global north. How does the context of development influence and change debates around ethical questions such as bias, fairness, and accountability? The exploration has given us a base on which we hope to continue building a global south-focused discourse, similar to what we have been doing now, for a while, with the use of data for development.

Much of our work over the years has followed technological development to understand their impact on the bottom of the pyramid and their potential for supporting development goals. We took a completely different turn, when we embarked on another new journey of looking ahead. In December 2019, we were commissioned to conduct a horizon scan to identify the megatrends that would affect the Asia Pacific region leading up to 2030. What we found is detailed in the Megatrends section of this Annual Report. The work was primarily a synthesis of research that was already conducted in the area by a variety of stakeholders including government bodies, management consulting firms, leading global economic publications and think tanks.

In May 2019,the Department of Census and Statistics (DCS) in Sri Lanka appointed a Committee for modernizing its activities. Sriganesh was appointed as an advisor to this committee. There was interest in using satellite imagery for crop yield prediction, to which we responded with a proposal. This built on work that was already being done in collaboration with DataSEARCH at the University of Moratuwa.. In October 2019, Sriganesh also detailed comments on the priorities identified by the appointed committee to modernize the department’s activities.

We continue to explore the use of publicly available remote sensing data and open-source analytical tools to understand agriculture, patterns of urbanization, and changes in land use and land cover. We also continue in our efforts to shape the discourse around data for development, through engaging with others who are interested in the field. In October, 2019, Sriganesh and Thavisha addressed the Global Shapers community of the World Economic Forum at the Shape of South Asia 2019 session on “Igniting innovation through data”.