Learn Python by Building Data Science Applications

Exploring unsupervised learning

First, let's try clustering our data. Clustering is a type of unsupervised learning with the goal of grouping records based solely on their features. It is often used to get a better understanding of the data before building a supervised model or as part of the exploratory analysis. It also could be used on its own. One common task is defining the target audience for the service or product. In our case, this should reveal the similarities between the battles across the dataset. 

This task may seem to be simple for a 1- or 2-dimensional (one- or two-column) datasets—indeed, our eyes and brains are splendid at finding clusters visually. It is, however, a near-impossible task for a human when the number of dimensions ...