In my role as vice-chair of the World Economic Forum’s Global Agenda Council on Data-Driven Development, I think a lot about how to use big data to improve lives. And because I’m confident in the power of analysis to uncover hidden relationships and spot emerging trends, I know that data analytics will play a major role in vanquishing Ebola.
Why am I so sure?
We need only look to the past. In 2003, after the outbreak of SARS, the Hong Kong Department of Health decided to prepare for the next pandemic with algorithms, by modernizing its analytics platform. With the ability to identify hotspots and forecast where the disease was likely to spread, the department fought an outbreak of Dengue fever with no lives lost.
But there’s another reason to be optimistic. In recent years, the technology game has changed. We can now process huge volumes of data – even unstructured data such as texts and other sentiment-based materials. And given advances in hardware, we now have enough affordable memory capacity to build the complex models required to answer previously unanswerable questions.
It’s a fundamental shift in approach. In data-driven analytics, information leads us to questions we haven’t even asked yet. In the newer, model-driven approach to analytics, we start with a question, and that leads us to the relevant data, which often resides outside conventional places. Creating models isn’t easy: it takes expertise and research. But once we have them, we can apply them to existing data to make breakthrough discoveries.
With Ebola, that’s already happening. Anonymized data from mobile phones points to regional population movements that can help predict where disease clusters might pop up next. My colleagues at the Swedish non-profit Flowminder built models that are helping officials in Senegal know where to focus preventive measures and healthcare.
In Sierra Leone, IBM has partnered with NGOs and telecommunications companies to develop a system that lets people report Ebola-related issues via phone and SMS. IBM is using the data to create opinion-based heat maps which correlate public sentiment to location, so officials know where supplies and other assistance are most needed.
And we can learn from other disasters, as well. In the Philippines, the International Organization for Migration (IOM) took a similar approach in responding to the destruction wrought by Typhoon Haiyan. Originally, the data was in Excel spreadsheets, but with the help of us at SAS, IOM transferred it to data-visualization software, and insights quickly emerged. As it turned out, what people needed was different from what IOM had assumed. Even more than food or medicine, people needed fuel – to run hospital generators and to transport people from badly hit islands to safer places. This is model-driven analytics at its finest.
I’m glad to see efforts like these, and I want to call on our industry to do all it can to support public-private partnerships even more, because it’s going to take all of us to turn the tide.
Securing public health requires new levels of connectivity, cooperation and open-platform collaboration. By combining traditional data sources with open-source intelligence captured from either new media or unconventional data sources, public-sector agencies are well placed to see disease as it emerges – and shift the focus from reacting to “predicting to prepare” or even “predicting to prevent”.
In the past, fighting an adversary such as Ebola would have been almost impossible, but given today’s new convergence of computing capacity, big data and modelling capability, we have the tools to get it done. All we need now is inclusive and open collaboration.
Author: Mikael Hagstrom is executive vice president at SAS, and vice chair of the Global Agenda Council on Data-Driven Development at the World Economic Forum.
Image: Doctors look at a scan of a patient on a computer screen at the Korle Bu Teaching Hospital in Accra, Ghana, April 24, 2012. REUTERS/Olivier Asselin