If we are talking about machine learning, why should we care about visualization? The answer is simple: if you cannot show what you have analyzed and the outcome of your models to somebody without any technical knowledge, then you will not be able to show any added value. We have already shown how important data visualization is for understanding a dataset and to decide which features will be most useful for training our model. We are now going to investigate which type of diagram is best suited to tell the story of our data and the new information we got from it.
The following topics will be covered in this chapter:
- Showing basic comparisons and relationships between variables
- Building data distributions using histograms
- Representing geographical distribution of data in maps
- Showing data that changes over time