The aim of this project is to work with messy and unbalanced data. We will use the "Census Income" dataset available from UCI Machine Learning Repository, which contains the income level of a group of people and a set of variables to describe each person. The objective is to try to predict if people will earn more or less than $50K.