Despite progress, about 300 million people in Africa still live in extreme poverty. The global community has agreed to execute the Sustainable Development Goals (SDG) to keep improving the human condition. However, due to a data challenge, researchers lack the information required to deepen our knowledge of the causes and consequences of poverty.
Our project aims to combine artificial intelligence methods and mainly publicly available satellite images to produce multidimensional poverty data for all communities in Africa from 1984 to 2020.
To address this aim, we will (i) train and adapt image recognition algorithms (CNNs) to identify multidimensional poverty from Landsat satellite images of African communities over time and space, quarterly, from 1984 to 2020. Using these CNNs, we will (ii) compare the quality of poverty data produced by Landsat images, Google Maps (year 2005-), Sentinal 2 (2015-), RapidEye (2009-2020) PlanetCope (2009-). We will (iii) create statistical software, OvervatoryOfPoverty, that enables us, and others, to access the poverty estimates from obj1-2 and to produce new estimates using different satellite technologies. Based on our methodological advances, we will (iv) forecast at what levels and speed African communities are lifting out from multidimensional poverty and their likelihood to achieve the relevant SDGs by 2030. We will conduct all of our satellite image processing via the Google Earth Engine (GEE). Consequently, as we will process a vast amount of image data, acquiring unrestricted amounts of data and GEE use, along with free technical support, will make a significant difference for our project. As an added benefit, we expect a fruitful intellectual exchange of ideas with the GEE community. These are the main reasons for our application to SNIC.