Stop Using the Elbow Method: Smarter Ways to Pick K in K-Means

The right K in K-means means more accurate insights, more substantial business impact, and more credibility when you present results.

Someone on Reddit recently posted:

I am working on a dataset that has 19 columns and 36000 rows...I am asked to perform clustering on it. So I am experimenting with KMeans. When performing the elbow method for this problem I am getting the following graph. The number of clusters were from 1 to 249. Can anyone suggest me the value of k which I should choo…