Data Visualization from Simple Plots to UMAP

Course Description

Visualizing data is often the key to understanding it. This course gives an overview of common data visualization techniques and considerations. Attention will given to exploratory visualization particular to chemometrics. We’ll explore the elements of different visual representations of data, the tools used to create plots, and ideas on how to enhance your visualizations. Both conventional and image data will be considered. Along the way we’ll see how Matlab and Eigenvector plotting tools work and how to use them. New methods for visualizing data will be introduced. The course includes hands-on computer time for participants to work example problems using PLS_Toolbox and MIA_Toolbox, or Solo+MIA.

Prerequisites

Chemometrics I — PCA, and Chemometrics II – Regression and PLS or equivalent experience.

Course Outline

1. Overview and Motivation
   1. Data Structures and Features
   2. Exploratory vs. Explanatory
   3. Practical Considerations
   4. Basic Plotting Elements
   5. Matlab and PLS_Toolbox Plotting Organization
2. Plot Types
   1. Popular types
   2. Images
3. Custom Plotting
   1. Defaults (Matlab, Plotgui)
   2. Custom Class Symbols
   3. Reverse Engineering
   4. Fonts
   5. Static Plots
   6. Movies
4. Exporting
   1. SVG vs Bitmap
   2. Publications
5. Higher Order Data
   1. Colorby
   2. Layering
   3. Tiled plots
   4. Condensing Data with Factor Based Models
   5. Slicing and Motion
6. Beyond PCA: New methods for Data Visualization
   1. t-SNE, t-distributed Stochastic Neighbor Embedding
   2. UMAP, Uniform Manifold Approximation and Projection
7. Conclusions
   1. Resources
   2. Future Considerations