CMU-R researchers and students at the forefront of the Big Data Revolution -Carnegie Mellon University Africa - Carnegie Mellon University

CMU-R researchers and students at the forefront of the Big Data Revolution

CMU-R researcher at the forefront of the Big Data revolution in Africa

With the open data policies and advances in technology, particularly the development of the Internet of things, we now live in a world overflowing with data. The devices that we use in our daily lives are constantly transmitting bits of information about where we are, what we are doing, and what we are purchasing. The main thrust of Big Data research in CMU-R is how to harness this fantastic resource to help us improve the way we live. 

Scientists have been working with "big data" for decades in research areas such as statistics, astrophysics, computational engineering: it is the application of big data in social sciences and human behavior that has caused the real buzz in recent times. 

How can we turn data into information, into insights and eventually actions that are truly useful for individuals, organizations and society? If we make accurate and useful predictive models from this readily available data, we can better plan our cities, our future, our lives and our everyday activities.

The applications of big data are multiple. At CMU-R we've looked at applications in Healthcare, Economics, Finance, Health, Agricultural, and Transport and we've even been able to prove that dancing can make you smarter!

Healthcare: Baby weights and real-time information on Malaria Prevalence

Our research in healthcare showed that health workers took much more care with weighing ne

wborns when required to input the weight digitally rather than on paper, showing that digitalization is not only more efficient with storing, searching and transferring data but will help improve the accuracy of information being collected. This work was recently published in the British Medical Journal. 

Again in health systems, big data can serve as an alternative view on important metrics, which may give a very different perspective on the reality of what is happening then other traditional data sets. For instance, in areas where malaria is present like Kigali, it is common for people to simply go to the pharmacy and purchase malaria medication and bypass medical clinics. We used Point of Sale data from pharmacies on the sale of Coartem, a malaria treatment to infer that the incidence of malaria in Kigali is driven by the rainy seasons and is trending upwards as confirmed by the Ministry of Health.  The monthly incidence rate generated by this approach can help to monitor malaria cases, inform policymakers and warn when we are most at risk from malaria.

Economics: Open Access to Consumer Prices and an Alternative View on Poverty

How can we measure poverty? Instead of attempting to measure poverty directly through surveys, we can use different sources of data to estimate levels of disposable income in various geographic areas within a country. We looked at satellite imagery for nightlights, cross-referenced with population density and mobile records (both for mobile ownership and call volume) to predict multidimensional poverty indices by sector in Rwanda. This work could form a real-time monthly poverty index by geographical location which would give governments and other organizations vital information on the progress of poverty reduction projects. Normally poverty indicators can only be gathered from census data which is costly and cannot be updated frequently.

CMU-R's Professor McSharry and his team built a dashboard for the Consumer Price Index calculated by the National Institute of Statistics Rwanda (NISR), which enables anyone, anywhere to monitor inflation, obtain information about prices of specific consumables and download the data. Rwanda is the first country in Africa to make official data publicly accessible in this manner and Professor McSharry hopes that it will encourage other countries to expand their open data policies. 

Planning: From weather effects on crops to mobile phone logs for city transport planning

We worked to use big data to help with planning and access to information. In collaboration with the Rwanda Regulatory Authority (RURA), we gained access to call records, from which we developed models on where people live and where they work to better plan city transport routes for Kigali.  In Europe, this is usually done by sending enumerators to talk to commuters, which is time consuming, biased and costly, involving cooperation with traffic police. Crunching mobile data is a better solution all around.

Our work on predicting the effect of weather on agricultural crops like tea will allow index insurance models for farmers to mitigate their climate risks. It also provides vital information for those planning where to grow tea in the future to be resilient to climate change. Global warming suggests an emphasis on planning for more high altitude tea where temperatures are lower and the quality of tea improves. 

Other applications: Can dancing make children smarter?

Finally, data analytics can be used to prove effectiveness in the NGO sector. Our student practicum project was undertaken with Mindleaps, who developed an innovative method to bring street children back into the formal education system by first engaging them with dance classes. A team of CMU-R students were able to show that the dance classes were improving the children's cognitive abilities, many of whom had never attended school, and allowed the Mindleaps team to select which students were ready for formal schooling. 

Professor Patrick McSharry, who is a leader in Big Data research, has been part of the team in Carnegie Mellon University in Rwanda for two years, as a visiting faculty from Oxford. He recommends Rwanda as a fantastic country to engage in research at the interface of data science and development. 

Related Links:

CMU-R Data Science on CNBC Africa

The Big Data Revoultion

MindLeaps Rwanda

Mindleaps Streetchildren Program

Professor Patrick McSharry