We aim at modelling 2017's french presidential election by using machine learning algorithms on different subsets of the INSEE databases. Our goal is to answer the following questions : how accurately can we predict the election results in a particular town, based on the several statistics the INSEE have produced once we have trained our model on a subset of french cities ? Which features have the greater impact on the results or on the turnout ? What can we learned about the candidates just by knowing the specificity and election results on each city ? In a more general way, we want to know in what extend the socio-economic factors are by themselves sufficient to predict the result of an election.
Please find the full report here : link