Mini 1, 6 units, Tuesday/Thursday 1:30pm-2:50pm

Wean 4709

This course will be a broad overview of determining and visualizing the structure in data. Basic graphical principles (e.g. Tufte) will be discussed, but the primary goal is to be able to utilize data analysis/visualization tools to explore data, summarize results, and identify structure, both expected and unexpected. Material will be pulled from books including Springer's "Handbook of Data Visualization" (Chen, Hardle, Unwin editors). This book in particular also contains several chapters on how to work with different languages/software packages. After some basics, topics will be chosen based on the research interests of the class.

basics of graphics, design choice, options, Tufte's Graph Principles

histograms, average shifted histograms

kernel density estimates

sploms, parallel coordinates, mosaic plots, trellis displays

3-d plots, projection pursuit, grand tour

multidimensional scaling, bipartite graphs

hierarchical trees, spanning trees, networks, directed graphs, treemaps

geometric graphs (delaunay triangulation, convex hull, etc)

comparing sequences

glyphs/icons

regression diagnostics, analysis of covariance plots, interaction plots, outlier detection

variable selection

smoothing techniques (nonparametric, splines)

clustering

visualizing matrices

social networks

visualizing Bayesian data analysis

There will be occasional homework and a final graphics design project. Complete understanding of the underlying statistical methodology is not a prerequisite but students should have a statistical background. The class will be appropriate for both beginning and advanced graduate students. Undergraduate enrollment requires instructor permission.