Assessment of multidimensional poverty alleviation effort in Tunisia

Long-term poverty data can support accurate decision-making by allowing decision-makers to track progress over time and identify areas in which additional resources or interventions may be required. In this study, we use the IPM to measure and assess poverty across regions and governorates in Tunisia, this index is a much more actionable and policy-relevant because it’s based on household survey information. However, such data, through its time-consuming, expensive, under-covered and labour-intensive, are available only once every ten years. Moreover, missing data is a common issue in survey datasets. Hence, we propose the deep learning method based on satellite images as a new and complementary approach to measure, evaluate and predict poverty in Tunisia.