Welcome to factor_analyzer Discussions!

[desilinguist]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[famarti]

Greetings! I'm Facundo from Argentina and i'm currently using the library to work with marine ice data. I had no issues so far, but I wanted to ask you if you know where I can find the equation used in the _varimax function (rotator class). I could notice that such function is used in all Varimax python libraries I found on the internet, but I can't find that kind of iteration in the books or papers. My intention is to quote your library, as well as the bibliography. PD: You did an awesome job!

[spider-z3r0]

Hello all,

I'm considering using this module to teach undergraduates about the use of EFA, but the output isn't really conducive to new learners. For example, It's not easy to make a scree plot or to clearly see which vars are loading onto which factor. Has anyone tried to 'humanise' the output from this otherwise good package? The docs aren't great for this.

[desilinguist]

Hi @spider-z3r0, you are right. This package wasn't really designed for learning factor analysis. Its primary goal was to take the excellent factor analysis capabilities available in R and make them available in Python for folks who have some experience with factor analysis. We'd be happily welcome any efforts to make the output more accessible to learners.

[spider-z3r0]

Thanks for the response. I understand your reply, but I think some work on the output would be really valuable. I'm happy to give it a shot but I'm not a competent enough python dev to do this efficiently. Would it be possible for me to ask some clarifying questions to make sure that the things I've tried are accurate?

[dsimon-vr]

Hello, apologies for what is probably a basic question: before running factor analysis, do my input features need to be standardised or should I use the raw values?

[DavidGuzzi]

the data does not need to be scaled (is performed by the algorithm)...

[DavidGuzzi]

Hello everyone.

First of all, many thanks for the effort in developing this library for all Python users. There's nothing better than open-source development.

For this reason, and as a recent user of the library, I would like to share a question that could potentially lead to an improvement for factor-analyzer.

While working simultaneously with Stata and Python libraries that offer Factor Analysis, I realized that the method Stata calls Principal Factor (one of the four possible methods, as supported by books like Rencher and Christensen, Methods of Multivariate Analysis (2012), Chapter 13) was not present in any version of Python. I'm particularly intrigued by the "principal" method in factor-analyzer, as it is not the Principal Factor method, but rather what both Stata and the literature refer to as Principal-Component Factor.

Could you please confirm if this is the case? What is the reason behind it? I believe the Principal Factor method is very interesting since the estimated Psi (Ψ) matrix is considered in the estimation of the covariance matrix or correlation matrix (which does not happen with the Principal-Component Factor method).

If needed, based on my reading of the literature and the factor-analyzer repo, I have developed a brief (and not very elegant) class that provides results according to the Principal Factor method, and its outputs are consistent with those from Stata.

I hope this is helpful. Once again, thank you very much. I'm available for any further questions.

Best regards,
DG.