Correspondence analysis

#1

Can Dundas BI do Correspondence analysis?

Has anyone done it?

Thanks

#2

Hi David. I can’t speak from experience, but there are packages in R and Python designed to perform these calculations, and it can then be output for visualizing in a scatter plot, etc.

#3

Thanks Jamie - I suspect I will have to go down that route.

I’ve never used a data cube python transform, nor bridge parameters before. It will be a learning experience for me, so anyone with experience who can give me some pointers is more than welcome to chip in!

#4

Also depends on what your Dundas server is running. If it’s Windows, only one person can use R at a time. Python doesn’t matter.

The Windows version of Rserve works in cooperative mode only, which means that only one connection at a time is allowed and all subsequent connections share the same namespace. For example, using both the R Data Generator transform and the R Language Analysis transform in the same data cube will result in an error because of this limitation.

https://www.dundas.com/support/learning/documentation/installation/how-to/install-and-configure-r

#5

Hi Christopher

Yes, we gave up on R a long time ago - the server kept falling over and not restarting. Definitely going down the Python route.

Seems that Dundas only support Python 3.6 and not the latest 3.7,

1 Like
#6

We also had a lot of issues with the windows version of R. We were able to keep things more stable by using the Linux version but i’m also a big fan of Python as an alternative too.