Chair for computeroriented statistics and data analysis

Visualising Information
Interactive graphics are a very effective and intuitive way of gaining insights from data. We have developed and are further developing these methods in the MANET package, especially for larger data sets, to provide Data Mining Visualisation software.

Large Data Sets
Large data sets are complex, there are far more date, far more variables and far more structure. They arise in environmental investigations, financial data bases and everywhere. The data can be analysed in much greater depth, but classical statistical methods do not always scale up. We are interested in more flexible data manipulation tools, better analytic tools and new visualisation methods. Visualisation is essential to get an overview of the data, to monitor data quality and to interpret results.

Exploratory Data Analysis and Exploratory Modelling Analysis
Modern, innovative software enables not only true Exploratory Data Analysis but also Exploratory Modelling Analysis, that is the reviewing and comparison of multiple competing models of data. MANET is primarily designed for EDA, but also gives an overview of many models through graphical analyses of residuals. TURNER offers many tools for exploratory investigation and manipulation of contingency tables, but its strength is the flexible comparison of loglinear models. It is a major aim of the department to extend the ideas underlying Exploratory Modelling Analysis.

Software development
Methods have to be theoretically researched, implemented in software and tried out in practice. Our group lays great store on the elegant, consistent and intuitive software implementation of our ideas.

Cooperation
We cooperate with other university researchers, with administrative organisations and with commercial companies. We are always interested in new applications for testing our ideas and for the generation of novel visualisation and analytical methods.