Unemployment is a social scourge that imposes costs on society far beyond financial ones. Unemployed individuals not only lose income but also face challenges to their physical and mental health, which translates into negative effects on families, relationships and communities. Therefore, we aim at delineating a strategy to understand which features of individuals (or their households) are most closely associated with the probability of being unemployed. Eventually, we want to be able to predict individual unemployment status using modelling techniques. For that, we will use data from the Current Population Survey (CPS) monthly survey, which provides information about key demographic and labor force characteristics. After performing the necessary transformations and feature selection techniques, we will perform a brief descriptive exploratory analysis. We then use three different types of models - Lasso, Logistic and Random Forest - to select relevant variables for predicting unemployment. Using data from the survey in 2018, we will evaluate the performance of our model and draw a series of final conclusions concerning model choice.