Thanks for the helpful write up on your process.

Points of feedback:

1) The dashboard would have been stronger if you'd named the five clusters based on the risk of fraudulent behavior.

2) It's not immediately apparent how the dashboard answers the question you set out in your objective statement. I actually would have started with view #3 and then used the other two as backup.

3) Since color is a primary tool for users to identify each cluster, the color scheme assigned to clusters should not be used elsewhere in the dashboard.

All that being said, this was a valuable exercise exploring a damning problem with the HC industry. I really appreciate the simplicity of your explanations in text and in the video!!

(For more Tableau tips, check out: https://towardsdatascience.com/new-data-science-f4eeee38d8f6#04d5)