A Little Book of Python for Multivariate Analysis

This booklet tells you how to use the Python ecosystem to carry out some simple multivariate analyses, with a focus on principal components analysis (PCA) and linear discriminant analysis (LDA).

The jupyter notebook can be found on its github repository.

Notes

This booklet assumes that the reader has some basic knowledge of multivariate analyses, and the principal focus of the booklet is not to explain multivariate analyses, but rather to explain how to carry out these analyses using Python.

The naming conventions in the functions were kept like in the original source for compliance. The variables were renamed to more generic names, so it would be possible to load your own dataset and run the notebook as it is if the first column of your data contains the data classes. See the cell that does the reading of the data.

The Python code was aimed to be easy to understand, like the R code in the original source, rather than be computationally and memory efficient.

Contents

• A Little Book of Python for Multivariate Analysis
  • Setting up the python environment
    • Install Python
    • Libraries
    • Importing the libraries
    • Python console
  • Reading Multivariate Analysis Data into Python
  • Plotting Multivariate Data
    • A Matrix Scatterplot
    • A Scatterplot with the Data Points Labelled by their Group
    • A Profile Plot
  • Calculating Summary Statistics for Multivariate Data
    • Means and Variances Per Group
    • Between-groups Variance and Within-groups Variance for a Variable
    • Between-groups Covariance and Within-groups Covariance for Two Variables
    • Calculating Correlations for Multivariate Data¶
    • Standardising Variables
  • Principal Component Analysis
    • Deciding How Many Principal Components to Retain
    • Loadings for the Principal Components
    • Scatterplots of the Principal Components
  • Linear Discriminant Analysis
    • Loadings for the Discriminant Functions
    • Separation Achieved by the Discriminant Functions
    • A Stacked Histogram of the LDA Values
    • Scatterplots of the Discriminant Functions
    • Allocation Rules and Misclassification Rate
      • The Python way
  • Links and Further Reading
  • Acknowledgements
  • Contact
  • License

License

A Little Book of Python for Multivariate Analysis by Yiannis Gatsoulis is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Based on a work at A Little Book of R for Multivariate Analysis by Avril Coghlan licensed under CC-BY-3.0.