36315 Statistical Graphics and Visualization, Spring 2002
Instructors:
Tom Minka, Statistics Dept, Baker Hall 228D, minka@stat.cmu.edu
William Eddy, Statistics Dept, Baker Hall 132F, bill@stat.cmu.edu
Teaching Assistant:
Fang Chen, Baker Hall A60D, fangc@stat.cmu.edu
Lectures: Monday and Wednesday, 12:301:20, Doherty 2105
Computer labs: Tuesday, 12:301:20, Baker 140c or Wean 5205
Overview
Graphs are the most thorough and persuasive method for transforming
data into knowledge and action. The aim of this course is to help you
turn passive numbers into powerful statements.
Realizing the potential of graphics requires methods and basic
principles. Data can be confusing, and a poorlychosen graph can lose
your audience's interest, hide the truth, or even lie. Statistical
graphics is not merely the cramming together of information or making
a novel kind of display. It is about effective communication: taking
a potentially complex and confusing message and turning it into
something readily and enjoyably understood.
Course Objectives
We will strive to teach you three things in this course:

How to design an effective presentation of data.

How to use visualization for exploratory data analysis.

How to produce statistical graphics with the software package Splus.
Relation to Data Mining
Data Mining (36350) is a companion course offered in the fall which
focuses on data analysis through visualization and modeling. It
covers far more methods than covered here and is more advanced,
requiring a greater knowledge of statistics. But 36350 does not
delve into graphical perception, maps, dynamic graphics, or interactive
graphics. The visualizations used in 36350 tend to be more technical
and specialized for the purpose of analysis.
Schedule
The schedule is organized around data of increasing dimension:
1D, 2D, 3D, and beyond.

univariate data

histograms, density estimates, boxplots, errorbars, quantiles, md plot

pies, bars, dotcharts

transformation, comparison, sorting, grouping

bivariate data, time series

scatterplots, loess, cut & stack

perception, display principles

color, jittering, banking

surfaces, response tables

level plots, profile plot, multiway dotchart

contingency tables

maps

dynamic graphics

multivariate data
Grading policy
The grade will be based 50% on homework and 50% on the project.
Homework each week will be to write up the result of your lab.
The project will be developed in pieces during the semester and due
during finals week.
Lab room assignments
Lab space is limited so the class will be split among two rooms.
If your last name starts with AL, go to Baker Hall 140c.
If your last name starts with MZ, go to Wean Hall 5205.