Visualizing the survival rate for each group of passengers provides us with an overview of the data. We have an idea about how the different features interact with the survival rate and with each other. For instance, we know that the social class has a different impact on the survival rate depending on the gender. But which of the two features has the highest impact? How big is the impact of each feature? We haven't defined a ranking of the features or quantified the impact of each of them. Some machine learning techniques allow us to investigate further, answering to our questions.
We have three features (class, sex, and age) and we want to divide the passengers in groups accordingly. We can't define a group for each combination of features since we won't have enough data for some groups such as the female children of the first class. A solution is to divide the passengers in groups in such a way that...