Ebola: Why Big Data Matters

Ebola: Why Big Data Matters
Ebola: Why Big Data Matters
The Ebola virus outbreak in West Africa has now claimed more than 4,000 lives and has entered the borders of the United States. While emergency response teams, hospitals, charities, and non-governmental organizations struggle to contain the virus, could big data analytics help?

A growing number of Data Scientists believe so.

If you recall the Cholera outbreak of Haiti in 2010 after the tragic earthquake, a joint research team from Karolinska Institute in Sweden and Columbia University in the US analyzed calling data from two million mobile phones on the Digicel Haiti network. This enabled the United Nations and other humanitarian agencies to understand population movements during the relief operations and during the subsequent cholera outbreak. They could allocate resources more efficiently and identify areas at increased risk of new cholera outbreaks.

Mobile phones, widely owned even in the poorest countries in Africa. Cell phones are also a rich source of data irrespective of which region where other reliable sources are sorely lacking. Senegal’s Orange Telecom provided Flowminder, a Swedish non-profit organization, with anonymized voice and text data from 150,000 mobile phones. Using this data, Flowminder drew up detailed maps of typical population movements in the region.

Today, authorities use this information to evaluate the best places to set up treatment centers, check-posts, and issue travel advisories in an attempt to contain the spread of the disease.

The first drawback is that this data is historic. Authorities really need to be able to map movements in real time especially since people’s movements tend to change during an epidemic.

The second drawback is, the scope of data provided by Orange Telecom is limited to a small region of West Africa.

Here is my recommendation to the Centers for Disease Control and Prevention (CDC):

  1. Increase the area for data collection to the entire region of Western Africa which covers over 2.1 million cell-phone subscribers.
  2. Collect mobile phone mast activity data to pinpoint where calls to helplines are mostly coming from, draw population heat maps, and population movement. A sharp increase in calls to a helpline is usually an early indicator of an outbreak.
  3. Overlay this data over censuses data to build up a richer picture.

The most positive impact we can have is to help emergency relief organizations and governments anticipate how a disease is likely to spread. Until now, they had to rely on anecdotal information, on-the-ground surveys, police, and hospital reports.

Comments