I love the UN Global Pulse. In a recent report “Big Data for Development,” the organization chronicles how Big Data—call logs, mobile-banking transactions, online user-generated content such as blog posts and Tweets, online searches, satellite images, etc.— can be turned into actionable information. The report reveals interesting ways that this data can be used:
- Discover “digital smoke signals”—anomalous changes in how communities access services, that may serve as proxy indicators of changes in underlying well-being.
- Real time effectiveness of policy actions
- Data aggregations can reveal trends in a visual manner – where words and actions may seem innocuous on their own, tools like word clouds and data visualizations can change the way we understand and interpret data
- Discover health crises or epidemics – keyword searches for flu symptoms inspired Google Flu Trends, and the CDC has become extremely effective in predicting epidemics / outbreaks.
The report offers a really interesting example of how behavioral changes and general population well-being are exemplified through collections of data:
(1) A local mobile operator may see many subscribers shift from adding an average denomination of $10 on their SIM-cards on the first day of the month to a pattern of only topping off $1 every few days; The data may also show a concomitant significant drop in calls and an increase in the use of text messages;
(2) Mobile banking service providers may notice that subscribers are depleting their mobile money savings accounts; A few weeks into this trend, there may be an increase in defaults on mobile repayments of microloans in larger numbers than ever before;
(3) The following month, the carrier-supported mobile trading network might record three times as many attempts to sell livestock as is typical for the season.
(4) Health hotlines might see increased volumes of calls reporting symptoms consistent with the health impacts of malnutrition and unsafe water sources;
(5) Other sources may also pick up changes consistent with the scenario laid out above. For example, the number of Tweets mentioning the difficulty to “afford food” might begin to rise. Newspapers may be publishing stories about rising infant mortality;
(6) Satellite imaging may show a decrease in the movement of cars and trucks travelling in and out of the city’s largest market;
(7) WFP might record that it serves twice as many meals a day than it did during the same period one year before. UNICEF also holds daily data that may indicate that school attendance has dropped.
While data will never replace human interaction and observation, it is undoubtedly an incredibly useful glimpse into the lives of people locally and globally. As we continue to use design to address development, I look forward to seeing how data can enhance the policy and program design process.
