Hands-On Data Analysis with Pandas

Grouping planets by orbit characteristics

When we plotted out the period versus semi-major axis for all the planets using a log scale for the period, we saw a nice separation of the planets along an arc. We are going to use k-means to find groups of planets with similar orbits along that arc.

As we discussed in the Preprocessing data section, we can build a pipeline to isolate the scaling of our data, culminating with the KMeans object to make eight clusters (for the number of planets in our solar system):

>>> from sklearn.cluster import KMeans>>> from sklearn.pipeline import Pipeline>>> from sklearn.preprocessing import StandardScaler>>> kmeans_pipeline = Pipeline([...     ('scale', StandardScaler()), ...     ('kmeans', KMeans(8, random_state=0)) ...