When conducting principal component analysis (PCA) using SPSS, we can obtain both the component matrix and the component plot in the output. Are these two the same thing? In other words, are the values of the plotted points in the component plot equal to the numerical values in the component matrix?

I find the component plot very intuitive, so I would like to recreate it for use in my research paper. Therefore, I want to understand where the data in this plot comes from.

------------------------------
Alice Fujiwara
------------------------------

If by component plot you mean the loading plot, then yes, the loading plot is the scatterplot of the loadings taken from the component loading matrix.

------------------------------
Kirill Orlov
------------------------------

Thank you!

But sorry, I am not sure what loading plot is.

I got component matrix and the component plot in the output of PCA (see figures attached).

I find the component plot very intuitive. However, I feel that the original figure may not be suitable for publishing. So I would like to recreate it for use in my paper. Therefore, I want to understand where the data in this plot comes from.

------------------------------
Alice Fujiwara
------------------------------

Principal Component Analysis (PCA) in SPSS provides two outputs: the component matrix and the component plot.

1. Component Matrix: It shows the correlation coefficients between original variables and principal components. High values indicate strong correlations.

2. Component Plot: It represents observations projected onto principal components in a lower-dimensional space. Data points' positions come from scores calculated using loadings from the component matrix.

To recreate the component plot, extract scores from PCA results and plot data points in a scatter plot or 3D plot, providing an intuitive visualization of data structure and patterns.

Thank you very much!

The component matrix is the loading matrix and the component plot is the loading plot. They give the same numbers.

You can manually edit the plot (for example, rotate it as you like, add spike projections, etc.) You may also copy-paste the matrix to a dataset and build a  scatterplot yourself.

Thank you very much!