SBAIC News

Famine Early Warning through Rapid Data Analysis: The FEWS NET Data Warehouse

Carla Murphy, Communications Director, Kimetrica

The need for speedy, robust, and authoritative data sharing to manage and coordinate humanitarian response could not be more evident. This February, the United Nations made its first declaration of famine in six years for parts of South Sudan. More than 100,000 face starvation and a million are on the brink of famine. In three other countries—Yemen, Nigeria, and Somalia—it is estimated that another 20 million are also at risk.

Over the past four years, Kimetrica developed and is currently expanding the data management, analysis and reporting system used by USAID’s Famine Early Warning Systems Network (FEWS NET). Since 1985 FEWS NET has been a world sentinel, providing early warning and analysis on food insecurity in more than 36 countries. The FEWS NET Data Warehouse (FDW) is its new, increasingly high-powered information backbone that gives analysts the right data and tools at the right time to forecast and help decision-makers respond to acute food emergencies.

Of the FDW’s many achievements, perhaps the most immediately-observed was the time (and expense) saved in inputting, sharing and analyzing price data on 109 commodities in 500 markets from over 50 partners.

“FDW turned 30 days’ work into three,” said Jenn Karuri. Based in Nairobi, Karuri is a Kimetrica help desk supervisor who, in 2014, helped train analysts from around the world and in four regional FEWS NET offices on the new system. “It used to take them months to compile data and do several analyses but after FDW, the change was automatic.”

Other benefits include: 80 percent less calls to the Kimetrica help desk concerning bulk uploading or data collection issues, enhanced collaboration between country and regional offices and USAID headquarters, and more reliable data.

Before the FDW, opportunity for human error was ample. Each national office obtained price data in different formats, and data was often entered by hand. Errant keystrokes or, say, Somali place names with multiple spellings were common culprits that delayed reporting. But with built-in verification checks, the FDW was designed to anticipate and reduce human error.

The FDW was also developed to work seamlessly in countries with unreliable Internet. Data can be downloaded in less than 10 minutes, which can then be used offline to populate customized graphs. In addition to price data sets, we’ve since added: crop and livestock production; trade of agricultural commodities; nutritional status; population statistics for each FEWS NET country; and emergency needs estimates. The FDW is flexible and handles a range of data types, from time series to cross-sectional. (In its current phase of development, the FDW project is perfecting the FEWS Data Explorer, a powerful dashboard that analysts can use to create online visualizations and easily do price averaging, calculate terms of trade and more.)

Moreover, developing countries can also take advantage of the FDW’s standardization and ease of use to share data, particularly between border countries.

“The ability for border countries like Ethiopia and Eritrea to share data has improved their ability to understand border markets and trade patterns,” said Frank Riely. He is a senior technical adviser at Kimetrica and the former chief of party and technical coordinator for the FDW project.

The beauty of the FDW is that it forecasts the promise of “big data,” which is to help public institutions better manage and widely disseminate valuable information. Further FDW development will be critical not only for global humanitarian programming, but eventually for in-country systems as well.

“In the future, FEWS will be able to create sophisticated analytical tools that national early warning systems will be able to link into,” Riely said. “That’s what’s super exciting.”

Founded in 2006 by a group of former early warning and humanitarian specialists, Kimetrica provides clients with technical support in data analytics, software solutions, and research. We currently manage USAID FEWS NET through a small business set-aside, providing technical assistance in data, methods and research.  Other clients include the World Bank, DFID, the EU and several UN agencies.