Multiple View Framework for Highly Structured Data Buchkatalog

At the Department of Child and Adolescent Neuropsychiatry, Medical University of Vienna, a psychotherapeutic study is taking place, where alternative (cognitive behavioural) therapeutic processes of anorectic girls are analysed. Within that study a huge amount of time-oriented data is collected, which need to be synchronized according to different constraints to be comparable. Classical statistical methods are not applicable in the absence of concrete hypotheses and can hardly capture the complex process of therapeutic interventions. In this master thesis a multiple view framework with a powerful data preprocessing part is designed. The data preparation provides filtering, aggregation of values with functions and the conflation of records with different dates of measurement. The framework is implemented in a tool called LinkVis which visualizes the data with Chernoff faces, scatter plots and parallel coordinates. The scenario-based evaluation demonstrates the power of the LinkVis tool on real world problems and shows the data complexity. The tool is still a prototype and it needs further improvements.