Big Data 4 Development

I was competitively selected to attend the Self-Organizing Conference on Machine Learning 2016 organized by OpenAI which was held in San Francisco on October 7-8, 2016. OpenAI is a non-profit artificial intelligence (AI) research company initiated by Elon Musk and top research scientists in AI and machine learning (ML) to promote safe and friendly AI. Since its inception in late 2015 the company has attracted top researchers and scientists from both industry and academia to work on most interesting problems of AI. The Self-Organizing Conference on Machine Learning 2016 was an experimental gathering that OpenAI organized for the first time to promote collaboration amongst AI/ML researchers, overcoming the overheads of a conventional academic conference. This is how they presented the motive behind organizing this event.
I was invited to speak to the staff of the Joint Research Centre of the European Commission in Seville last Tuesday (11th October 2016). Their colleagues from Ispra, Italy joined in via video conference as well. I talked about LIRNEasia’s experiences and lessons in leveraging big data for public purposes. The slides that I used are available HERE.
We have been engaging with local universities from the start of our big data work, not just to source researchers and collaborators, but also to broaden the horizons of students. That big data can be leveraged for public purposes is not something that they had previously thought of till we arrived on the scene. This week (18th October 2016) we continued those efforts, conducting a lecture for students at the University of Sri Jayewardenepura on our ongoing big data for development research. The slides are available HERE.
LIRNEasia in partnership with the Centre for Internet and Society (CIS) convened a two-day workshop to discuss a ‘research and policy agenda on big data for sustainable development in the Global South.’ The workshop held in Madrid on the 8th and 9th of October 2016, was a side event of the International Open Data Conference 2016.  The objective of the workshop was to brainstorm ways of establishing Southern-led network to tackle some of the emerging opportunities and challenges in the use of big data in developing countries. The workshop explored a variety of issues around leveraging big data to tackle sustainable development. These include issues around representativity and marginalization, researching harms (competition, privacy, surveillance), researching solutions (legislation, regulation, ethics), and addressing challenges in relation to developing research capacity, accessing data and influencing policy.
Ideally, we would have had findings. But we are in the middle of research, so what we can present is work in progress: problems that have been faced; those that have been solved; those we’re still working on; etc. Hopefully, once we get our hands on the needed epidemiological data we will present findings in a few months. We are grateful to the incoming President of the Commonwealth Medical Association, Professor Vajira Dissanayake, for creating this opportunity for us. The presentation was made at a session chaired by Dr Hasitha Tissera, the Head of the Epidemiology Unit of the Ministry of Health.
I hope to write more about the insightful discussions at the workshop convened by LIRNEasia and CIS. For now, here are the slides I used to frame the discussion on Harms from Surveillance, (In)security, and impacts upon Privacy and Competition. Image source.
With support from the International Development Research Centre of Canada, LIRNEasia and the Center for Internet and Society are today convening a meeting of researchers working on aspects of big data for development in the Global South. The hope is that we will be able to contribute to shaping a research and policy agenda and map out a path for productive collaboration. A document was prepared as the basis for discussion. Here is an excerpt: However, it is important that those engaged in policy analysis make the effort to understand what data is available, in what formats and what is being done with it. For example, the mobile networks in developing countries are different from those in developed economies.
One of our current priorities is to work with the National Statistical Agency to see how we can complement official statistics instruments on socio-economic monitoring. China watchers are not collaborating with the NSO. They are trying to second guess it. I was wondering though, wouldn’t we be happy if we got these kinds of correlations? “Big data provide an increasingly comprehensive and timely lens” on the world’s second-largest economy, the analysts wrote, adding a caveat that such indicators should be interpreted with caution.
There appeared to be a problem with loading the slideset, so I went to Plan B. I was just about to do a big data talk with no slides. That is the first learning: always have a Plan B and be ready to improvise. This being Oxford, I thought they could access the slides off the Internet. But then the technical problem was solved and I gave a conventional talk.

Economics meets data science

Posted on September 5, 2016  /  0 Comments

In our big data for development work, we collaborate with data-savvy economists as well as economists who can code. Within Sri Lanka, we have not found them, but we keep looking. But looks like this is the future of economics. But what the tech economists are doing is different: Instead of thinking about national or global trends, they are studying the data trails of consumer behavior to help digital companies make smart decisions that strengthen their online marketplaces in areas like advertising, movies, music, travel and lodging. Tech outfits including giants like Amazon, Facebook, Google and Microsoft and up-and-comers like Airbnb and Uber hope that sort of improved efficiency means more profit.

Big data and agriculture

Posted on September 1, 2016  /  0 Comments

I was asked to make one point about the way forward at the closing session of the excellent e agriculture solutions forum organized by the FAO and ITU offices in Bangkok. Here is what I said (more or less, but this is the jist): Big data in agriculture We have come a long way from being fixated on radio as the be all and end all of ICTs in agriculture. We are fortunate to be living in an age when we can even take smartphones for granted in Myanmar, a country still listed as an LDC and one which went from 10 mobile connections per 100 people to over 80 in less than two years. Our own surveys (early 2015) showed that 63 percent of all mobile owners in Myanmar had smartphones, with more computing power than the computers we used just a decade ago. The mean price of a handset was USD87, with the largest number being in the USD 50 range.
As part of our big data work, we have been looking at sources other than mobile network big data for socio-economic monitoring. Night lights images from satellites was the favorite. But I’ve been always skeptical, partly because of looking down from the New Delhi-Colombo flight which flies through the middle of India and then right down the island from Jaffna. The intensity of the lights is so much higher in southern India, than in Sri Lanka. This story is about daytime images taken by satellites.
Lokanathan, S., Kreindler, G., de Silva, N. D., Miyauchi, Y.
In our article published last year on big data for urban development in the developing world, we said At one extreme of smart-city initiatives lies the vision of a centrally coordinated city resting on pervasive use of specialized sensors (e.g., one under each parking space; multiple sensors at intersections), real-time or non-real-time analysis of the resultant big-data flows, and reliance on mathematical models. South Korea’s Songdo is the exemplar. Reports of plans for green-field developments indicate that the Modi government is leaning toward this vision.
In our teleuse surveys, we found that missed calls beat out texts in some countries (e.g., Bangladesh v Philippines). One explanation is that there are more illiterate people in the countries where missed calls predominate. Qualitative research found that a lot of texting did not involve literate work, mostly it was forwarding messages sent by others.
LIRNEasia is currently hosting  Dr Ayumi Arai from the University of Tokyo’s Center for Spatial Information Science. She is also a Research Fellow with LIRNEasia collaborating on our big data for development research in Sri Lanka. We took the opportunity to organize a lecture for her yesterday (14th July 2016) for the senior staff of the Department of Census and Statistics (DCS) Sri Lanka, as preamble to a longer discussion with the department to collaborate with LIRNEasia and our partners on big data and official statistics in Sri Lanka. Dr Arai’s talk was on her ongoing Dynamic Census research work in Bangladesh which utilizes mobile network big data and official statistics to provide spatio-temporal insights on the socio-economic and demographic characteristics of the population at high granularity and high frequency. The slides from her talk are available HERE.