Monday, 22 August 2016

Fighting poverty with satellites

Governments in poor countries have to make important decisions to help their countries out of poverty. For these decisions, they need a lot of data, which is usually limited in those countries. Neal Jean and his team have found a solution.

No census causes old data
Surveys conducted in Africa in
the past few years

In poorer countries, it’s really hard to get detailed census data since roads and other infrastructure are generally not good so people are difficult to reach. On top of that, poor countries usually have lower literacy rates too. These two things, and other factors too, makes gathering detailed information about the people in that poor country especially hard. In Angola for example, 44 years elapsed since the last census before a recent one was conducted. In that time, the population grew from 5.6 million to 24.3 million and the country suffered a civil war. So governments have based their policies on horrible outdated data for an incredible long time. That’s mainly why this information is so important, especially there, since governments and policy-makers desperately need such information to make their countries better. Now, Neal Jean and his team have found a way to gather census-like data without actually conducting such a survey, but through the use of satellites.

Is it dark or really dark?
Scientists have already used night-time satellite images to determine which regions of the world are rich or poor in the past. This seems to be a really practical method; rich regions appear bright in the picture and poorer regions appear less bright. This happens because there’s less access to electricity there and also less artificial light. However, a disadvantage to this method is that it’s hard to see the difference between poor and very poor regions. They both appear equally dark. Scientists have also tried gathering information from smartphones, which are also sold more and more in poorer countries. Scientists can quite easily find out how wealthy somebody is based on his or her mobile phone use. The problem with this is that you can’t include the poorest of the poorest, which will cause wrong data. On top of that, most mobile phone data is owned by providers, and they aren’t that keen on giving their information away. Neal Jean and his team have found a method that doesn’t face either of those problems.

Combining day and night
They have designed software that can combine night-time images and day-time images. This method eliminates the problems you face when you use only night-time images. Because their software can recognize patterns that indicate wealth and poverty on the detailed day-time images. The software is then able to link the brightness of a place to its wealth or poverty levels. It can, for example, link large villas with swimming pools to bright light and crappy sheds to darkness. With the help of its knowledge, the software can then also tell the poor and the very poor regions apart, which is impossible when you only use night images. Another advantage they have is that Jean’s software only uses images and data that’s already freely available, unlike the mobile phone data. In other words, their software can be used as a replacement for surveys in countries where it’s hard or even impossible to conduct them. And it also shows how technological advances can also help fighting poverty. Or as economist Sendhil Mullainathan put it: “Why should the financial services industry, where mere dollars are at stake, be using more advanced technologies than the aid industry, where human life is at stake?”


Click here to read more about computer science.
Click here to read more about geography.

Sources: