A Matlab Toolbox for Data Exploration and Analysis, based on a data table class.
The toolbox is organized around a Table class, similar to the dataframe in R, which encapsulates data array together with row names, column names, table name. It also supports factor columns, such as categorical factors ("yes", "no"). Several utility methods are provided, and many plot functions have been override to automatically annotate the resulting plots with table meta-data when appropriate.
The toolbox also contains facilities for statistical analysis of data tables, such as principal component analysis, analysis of variance, or linear discriminant analysis. Again, intuitive methods are povided for exploring and analysing the results by taking into account the names of the rows or of the columns.
A presentation of the library is provided in the matStats-manual.pdf document.
Matlab provides a packaging mechanism via Add-On. The MatStats Add-On may be installed via the following:
- Seelct the "Add-Ons" tab from Matlab's main interface
- Choose "Get Add-Ons...". This opens the Add-On Explorer window
- Type "MatStats" in the search bar, and select the MatStats library.
- Click the "Add" button to add the library to your local configuration
To install the toolbox:
- download the zip archive of the latest release
- unzip the file
- add the 'matStats' directory to the Matlab path (see the "addpath" function from Matlab).
Some demontration scripts are provided in the "demo" directory.
The library has few (optional) dependencies:
- Some of the functions require the Statistics and machine learning toolbox from Matlab, but most of the features do not need it.
- The TableGui interface requires the GUI Layout Toolbox
- The MatGeom library is required for computing / displaying some features like inertia ellipses or ellipsoids.
The Matgeom dependency is automatically installed via the Add-On installation. Otherwise, dependencies need to be installed manually.