Ready for Anything: Disaster and Analytics

disaster analytics
by Derek Slater
Writer
, Ready State

Using data analytics, a humanitarian organization determines the most efficient ways to deliver aid to people affected by violence and natural disasters.

Outbreaks of violence are often unpredictable, creating extreme challenges for humanitarian aid organizations in providing safe shelter, food, water and medical care. That’s where analytics can help in predicting resources needed for people displaced by disasters.

Since 1951, the International Organization on Migration (IOM) has advocated for the well-being and dignity of migrants affected by violence and natural disasters. Today, the organization uses data analytics to determine the most efficient way to deliver resources to people in need.

The IOM works on the frontlines when worldwide humanitarian crises hit. For example, the Boko Haram conflict affecting the Lake Chad Basin area, particularly in northeast Nigeria, spawned large scale displacement and a significant humanitarian crisis.

According to the IOM, in Nigeria alone, more than 2.2 million people have been forced to flee their homes, and many are now living in temporary sites without basic services.

“When people are displaced, our ability to locate them and assess their needs is often limited,” said Nuno Nunes, a coordinator based at the IOM’s headquarters in Geneva, Switzerland.

Outbreaks of violence and natural disasters are hard to predict and dangerous, said Nunes, which makes delivering the right aid to the right places a challenge.

Today Nunes and his team collect data from hundreds of groups and thousands of individuals affected by disaster. They analyze past and present data — descriptive analytics — to understand how many people are on the move, where they are headed and what aid they need most.

In the future, Nunes says the goal is to go beyond descriptive to predictive analytics, which will help the IOM more reliably anticipate humanitarian aid logistics, such as the routes migrating populations will travel. The organization will then be able to deliver aid even more quickly, accurately and efficiently.

“Predicting what’s happening now, but inferring what you can’t directly measure can be extremely valuable,” said Bob Rogers, chief data scientist of analytics and artificial intelligence at Intel. “We’re always trying to extrapolate between what we can measure and what we need to know — both now and in the future.”

Swimming in Data

Analytics of any sort starts with data.

The IOM started in the early 1950s to help refugees displaced after World War II. The organization’s access to data has grown dramatically over the past decade, but it’s been a challenge to manage disjointed data systems.

Nunes said East Timor is a good illustration. In 2002, the Southeast Asian island nation gained United Nations status and, for several years, the construction of new infrastructure brought the population hope. When violent conflict erupted in 2006, however, many people were driven from their homes.

Nunes helped manage camps for displaced citizens. At that time, aid workers used registration systems to track the movement of individuals and the size of each relocation camp. The data helped track where to locate camps and supplies, but Nunes recognized the need for a global system that could better manage data collected from many different sources.

“Since 2010, together with a network of colleagues in emergency operations, we managed to further develop systems for tracking mobility, displacement and assistance needs, with a global footprint,” he said.

Today the IOM conducts regular, large-scale surveys in more than 30 crisis-affected countries, with approximately 2,700 census-taking organizations (including humanitarian and government partners) collecting data.

disaster analytics
IOM uses a database to track the movement of migrants worldwide, including South Sudan. Image courtesy of IOM.

This data helps analysts understand the situation on the ground in each camp location. However, the impact goes beyond tracking people, food and water supplies. For example, analytics can help catch outbreaks of malaria or polio more quickly, identify where to find the necessary medical supplies and send those supplies along the safest route.

Migration is an inherently complex problem, and with so many data points, countries and sources to collate, the IOM has a lot of data to analyze. Even as technology advances, human analysts continue to play an indispensable role.

The Path to Prediction

While the IOM’s analytics already help save lives today, Nunes and his team have set their sights on predictive analytics to deliver even more value tomorrow.

For the IOM, predictive analytics means accurately anticipating needs before they arise — for example, imagine correctly pre-empting a malaria outbreak before the first case is diagnosed.

“The way you model different diseases and how they spread, and the way people connect within an affected community actually profoundly influences what your strategy is for stopping the outbreak,” said Intel’s Rogers.

Rogers explained how predictive analytics looks at the connector points among data. In the case of the spread of disease, analytics can measure how the infected population connects and interacts with others in the community. Using this data, disease prevention strategies may include inoculation or isolating infected communities, he said.

In addition to delivering aid directly, the IOM helps set national and international policies around migrant populations. More evidence and more accurate foresight leads to better policy recommendations, and less reactive decision-making.

Nunes and his team are using the data to build statistical models for both disaster impact and migration routes. By understanding both, the IOM is better able to anticipate where camps and supplies should be located in order to best meet migrant populations’ needs.

“There will never be 100 percent accurate prediction; emergency contexts are volatile and dynamic by nature, and no model can fully account for all variables,” he said. But being able to look at probabilities with confidence makes the IOM’s work more efficient.

Still, achieving truly predictive analytics will demand even more data and more real-time analysis tools.

Rogers said leveraging Internet of Things (IOT) technologies, like putting location-tracking radio-frequency identification (RFID) tags on cell phones or wearable cameras, can help gather migration data.

“When you can go back over and re-analyze natural data with your new capabilities, you find all kinds of incredible information,” said Rogers.

The IOM continues to broaden its toolset for collecting and analyzing data, including open-source tools for mobile data collection using phones and tablets, satellite imagery for a broad overview of affected areas, and drones for a more detailed view of disaster impact and damages. IoT sensors also could play an increased role in capturing reliable data along common migratory routes.

Data capture through mobile devices will continue to grow with the introduction of IOM’s MigApp, which migrants can download for free to access emergency alerts, health information and services to assist during their migration crisis. In turn, IOM will capture the data to analyze global, regional and local migration patterns and trends.

The more information it has to analyze, Nunes said, the more IOM will be able to move to “real identification of problems and definition of new solutions, not only justifying the ones that already exist.”

Share This Article