K-centroids cluster analysis

[refmd]

Dear Community,

I have HR data, where I have to create different segments of employees to study each one and determine its financial effect on the organization. 

I'm trying to create a k-means clustering, but I have the following concerns & issues:

• I am building the model using 3 numerical variables (employee_age,  avg_monthly_salary, months_of_service), I'm not sure if these variables could be used to create the clusters or not. 
• What is the ideal way to visualize these clusters together so I can notice and compare if the clusters are really unique or they are overlapping, the below screenshot shows a sample of what I need:

I'm ok with using python or alteryx, so kindly any reference, comment, or note would be highly valuable. 

Thank you,, 

[PhilipMannering · Accepted solution]

I await better solutions, but one way could be to use plotly express' 3d chart.

It's still not that obvious where points are overlapping, but you can isolate each trace by clicking on the legend. See attached.

[PhilipMannering]

Documentation for this can be found here: https://plotly.com/python/3d-scatter-plots/ 

[refmd]

Many thanks, will try it 

[CharlieS]

Hi @refmd 

As a leader in the Alteryx Community, I have the ability to identify & mark accepted solutions on behalf of community members - and recently did so on this thread. If you have any questions or concerns with the solution(s) I selected please let me know by replying to this post.

Learn more about Accepted Solutions here.

Happy Alteryx-ing!