Clustering improvement needs

[rileyhales]

Faster run times - MiniBatchKMeans with lots of initializations? Standard Kmeans?

More figures

• plots of just cluster centers. grid of subplots, 1 per num clusters iteration. plot centers of all clusters on the subplot
• silhouettes
• inertia figure with knee labeled
• inertia and silhouette score on the same figure with separate y-axes

Compartmentalize plotting and summarizing functions so they can be run individually

Training data should be FDC from 0-100 in 2.5% increments -> 41 total points, round values to 3 decimals