Today, I learned a little more about clustering, specifically a concept called the “Elbow Method.” To figure out how many clusters we need I plotted this graph called WCSS, which is a way to see how close the points in a group are to each other. It helped me to figure out the best number of groups where adding another group doesn’t help much, we call this point the “elbow”.

Essentially WCSS (Within cluster sum of squares) is a measure of how tidy our clusters are. Adding up all the squares of the distances between each point and the center of the cluster. Lower WCSS give us neater clusters. As we increase the number of clusters, WCSS usually go down. We have the elbow method to find just the right number of clusters.