Hands-On Data Analysis with Pandas

k-means

One common clustering algorithm is k-means, which iteratively assigns points to the nearest group using distance from the centroid of the group (center point), making k groups. Since this model uses distance calculations, it is imperative that we understand the effect scale will have on our results beforehand; we can then decide which columns, if any, to scale.

There are many ways to measure distance between points in space. Often, Euclidean distance, or straight-line distance, is the default; however, another common one is Manhattan distance, which can be thought of as city-block distance.

Other clustering algorithms offered by scikit-learn can be found in the cluster module at https://scikit-learn.org/stable/modules/classes.html#module-sklearn.cluster ...