Abstract

The course will give an overview of developments in statistical graphics in the field of categorical data analysis, explain the concept of interactivity and describe modern methods of graphical analysis. The course will start from the basics of statistical graphics, so previous knowledge is not essential. For the second half of the course experience with statistical software and data analysis would be valuable. Throughout the course various software packages from the area of interactive graphics will be demonstrated, including Data Desk, JMP and our research software, MANET.

Objectives

Review graphic displays and interactive methods for categorical data. Introduce interactive graphics tools for categorical data.

Outline

• Visualising categorical data
  Barcharts, Spineplots (Piecharts etc)
• Introduction to interactive graphical methods
  Querying, Linking, Highlighting, Cueing
• Biplots (for correspondence analysis)
• Mosaic Plots for multidimensional data
  Construction, Interpretation, Variations
• Visualizing Loglinear Models

Target population

People interested in exploring and analysing data sets with categorical variables. The data sets could be small, multivariate studies or large surveys. They could also be the results of other analyses, such as the subgroups identified by data mining algorithms.

Learning outcomes

At the end of the course the participants will have

• a broad overview of visualisation methods for categorical data
• experience of exploring categorical data sets
• knowledge of the newest interactive methods for multivariate categorical data

Heike Hofmann, hofmann@vento.math.uni-augsburg.de
Antony Unwin, unwin@uni-augsburg.de
Department of Computer-Oriented Statistics and Data Analysis, University of Augsburg

The instructors are co-designers of the software MANET; Heike Hofmann is the author.

Maximum Enrollment: 30

Handouts

Spiral bound copies of all slides will be provided. Software and data sets will also be made available.