In this chapter, we explored different methods of dealing with missing data and learned how to group or summarize it. We have shown you how important it is to visualize the data after cleaning, in order to be able to understand and interpret the results, from basic to more advanced model predictions. This is the beginning of any feature engineering, since we transform and/or discard features based on their values. Too many missing values will imply that we cannot use that variable (or feature), or a high correlation will imply that we can discard one of the correlated variables. We will dive deeper into correlations in the next chapter, showing you how to measure them quantitatively, using different methods.
Preliminary data visualization is extremely important to gain an understanding of data properties and to interpret the results we obtain, even after applying a machine...
