Research Projects at CVAST

For a general description about research in Visual Analytics please see here at the overview page

In this project we support medical experts in exploring patients' conditions as well as the effects of clinical actions on patients' conditions.

Visual Analytics strongly emphasizes the importance of interaction. However, until now, interaction is only sparingly treated as subject matter on its own. How and why interactivity is beneficial to gain insight and make decisions is mostly left in the dark. Due to this lack of initial direction, it seems important to make further attempts in facilitating a deeper understanding of the concept of interactivity. Therefore, different perspectives towards interactivity as well as cognitive theories and models are investigated. The main aim of this research is to broaden the view on interaction and contribute to advance the field towards a sound theoretical grounding.

Advanced debuggers available in modern IDEs make it easier to debug programs and understand their runtime behavior. In this project we aim to enrich software debuggers in popular Java IDEs with visual methods that provide more insight into information available at the runtime. Example for this information are traces collected at tracepoints set by the user, the values of the variables over time, or values in large arrays.

Our efforts gave rise to the 'Eclipse Debugging Aids', a collection of free open source debugging tools for Eclipse.

Homogeneous multivariate data encompass multiple variables that have the same semantics. As example, these variables can represent the probabilities for a sample to belong to different classes, or item memberships of multiple sets.

With a large number of items, such homogeneous data tables become very rich of information that explains how the row entities are related to the different column variables, and how the columns are related to each other according to their relationships with the rows.

This project aims to develop visualization methods for analyzing homogeneous multivariate data. These methods should allow analyzing and selecting the row entities based on their relations with the different columns. Moreover, they emphasizes the column variables and the relations between them as the central part of the visualization, and allows analyzing these relations based on the row entities defining them.

Poor data quality leads to unreliable results of any kind of data processing and has profound economic impact. Although there are tools to help users with the task of data cleansing, support for dealing with the specifics of time-oriented data is rather poor. However, the time dimension has very specific characteristics which introduce quality problems, that are different from other kinds of data. To this end we tackle this important topic with Visual Analytics methods.

Data quality control can be divided into

  1. Data Profiling: identifying and communicating quality problems (e.g., w.r.t. specific Data Quality Metrics)
  2. Data Wrangling: transforming table formats or merging different sources
  3. Data Cleansing: correcting the found quality problems

Recent Publications