Associations between Pairs of Variables

Your data science project might not always include advanced algorithms. Sometimes, all you need is finding associations between pairs of variables. In addition, these associations might also be a preliminary analysis for advanced methods. There are many methods for measuring the associations, depending also on the type of the variables – discrete or/and continuous ones. This session introduces covariance and correlation between two continuous variables, contingency tables and chi-squared test for measuring the association between pairs of discrete variables, and T-test and analysis of variance for finding an association between a discrete and a continuous variable. The session introduces a bit more advanced methods at the end, namely Bayesian inference and linear models. All associations are also presented graphically. The demos are in T-SQL, R, and Python.