News

Lectures Mondays 10.00 - 11.30  in 3029 Wednesdays 10.00 - 11.30  in 3029 Tutorial Tuesdays 10.00 - 11.30  3029

Description

Graphical analyses complement statistical analyses.  Their advantages lie in checking data quality, in generating hypotheses (in contrast to testing them) and in explaining model results.  Graphical methods are becoming increasingly important for the analysis of large heterogenous datasets, for which classical model assumptions seldom hold.  In this lecture we will discuss graphical methods using examples from real applications, including some from the web.  R will be used primarily (especially ggplot2) with Mondrian for interactive graphics.

Literature

"Grammar of Graphics" Lee Wilkinson (Springer) "ggplot2" Hadley Wickham (Springer) "Interactive Graphics for Data Analysis" Martin Theus and Simon Urbanek (CRC Press) "Graphics of Large Datasets" Antony Unwin, Martin Theus and Heike Hofmann (Springer) "Handbook of Data Visualization" (eds. Chen, Härdle, Unwin) (Springer)

Software

R and Mondrian will be used in this course.

Datasets (R/Mondrian format)

Titanic Sparpaket

R-Code

UnivariateContinuous MultivariateContinuous CategoricalData SimulatingPlotsTitanic