Economic Well-Being Prediction Challenge
2000 Points
Can you predict a measure of wealth for different locations across Africa?
328 data scientists enrolled, 117 on the leaderboard
16 April—15 August
Ends in 20 days

Granular information on economic well-being is extremely useful for governments, policy makers, and NGOs. But household surveys that capture this kind of information are expensive and conducted infrequently in many African countries. For this competition we will attempt to create a workaround for this lack of data by building a model able to predict a measure of wealth as measured in household surveys using readily available inputs.

Using data from 18 different countries collected at different times, you must correctly predict the cluster-level estimated wealth measures found from surveys in 7 different countries not covered in the training data. A successful model could be useful for filling in the gaps between the more expensive surveys.