Practical Data Analysis Cookbook

Exploring correlations between features

A correlation coefficient between two variables measures the degree of relationship between them. If the coefficient is equal to 1, we then say that the variables are perfectly correlated, if it is -1, then we conclude that the second variable is perfectly inversely correlated with the first one. The coefficient of 0 means that no measurable relationship exists between the two variables.

Note

We need to stress one fundamental truth here: one cannot conclude that simply if two variables are correlated, there exists a causal relationship between them. For more information, refer to the following website: https://web.cn.edu/kwheeler/logic_causation.html.

Getting ready

To execute this recipe, you need pandas. No ...