Subscribe Bookmark RSS Feed

Clustering Twitter Accounts By Common Followers in JMP

rightmirec

New Contributor

Joined:

Mar 6, 2017

Hello,

 

I am looking to see if I can use JMP Pro to run a cluster analysis on approximately 40 social accounts and their followers. Ideally, I'd like to cluster the 40 accounts based on common followers. So far, I have sourced follower handles using FollowerWonk, and have them set up in a spreadsheet where each column represents a parent account (one of the 40), and their followers' handles are listed below them.

 

I have tried a basic hierarchal cluster, but the output was a black box. I'm a little lost on if JMP can even be used for this, so any help would be much appreciated.

 

Currently my data is formatted as follows:

 

 

Parent Account 1  Parent Account 2
@follower1  @follower2
@follower2  @follower3
3 REPLIES
Peter_Bartell

Joined:

Jun 5, 2014

Maybe if you share the analysis report we can provide some insight on what it may be telling you. There's no fundamental reason why you can't use this approach. You may also want to look at latent class analysis as well?

rightmirec

New Contributor

Joined:

Mar 6, 2017

Thank you so much for the quick response. I think my first error was running the cluster with the parent accounts as the column header, that left me with a dendrogram that looke like a black box. 

 

Hierarchical Cluster_Not Transposed.png

 

However, when I transposed my dataset (attached), I got a slightly more coherent analysis. I am still having trouble interpreting exactly what it means though.


Transposed Analysis.png

rightmirec

New Contributor

Joined:

Mar 6, 2017

Sorry, transposed dataset attached.