[Truncated abstract] Many databases in scientific and business applications have grown exponentially in size in recent years. Accessing and using databases is no longer a specialized activity as more and more ordinary users without any specialized knowledge are trying to gain information from databases. Both expert and ordinary users face significant challenges in understanding the information stored in databases. The databases are so large in most cases that it is impossible to gain useful information by inspecting data tables, which are the most common form of storing data in relational databases. Visualization has emerged as one of the most important techniques for exploring data stored in large databases. Appropriate visualization techniques can reveal trends, correlations and associations in data that are very difficult to understand from a textual representation of the data. This thesis presents several new frameworks for data visualization and visual data mining. The first technique, VisEx, is useful for visual exploration of large multi-attribute datasets and especially for exploring the correlations among the attributes in such datasets. Most previous visualization techniques can display correlations among two or three attributes at a time without excessive screen clutter. ... Although many algorithms for mining association rules have been researched extensively, they do not incorporate users in the process and most of them generate a large number of association rules. It is quite often difficult for the user to analyze a large number of rules to identify a small subset of rules that is of importance to the user. In this thesis I present a framework for the user to interactively mine association rules visually. Another challenging task in data mining is to understand the correlations among the mined association rules. It is often difficult to identify a relevant subset of association rules from a large number of mined rules. A further contribution of this thesis is a simple framework in the VisAR system that allows the user to explore a large number of association rules visually. A variety of businesses have adopted new technologies for storing large amounts of data. Analysis of historical data quite often offers new insights into business processes that may increase productivity and profit. On-line analytical processing (OLAP) has become a powerful tool for business analysts to explore historical data. Effective visualization techniques are very important for supporting OLAP technology. A new technique for the visual exploration of OLAP data cubes is also presented in this thesis.

Thesis (Ph. D.)--University of Oregon, 2008. / Typescript. Includes vita and abstract. Includes bibliographical references (leaves 147-151). Also available online in ProQuest, free to University of Oregon users.

Thesis (M.Sc.) - Carleton University, 2005. / Includes bibliographical references (p. 138-145). Also available in electronic format on the Internet.

Habilitation - Universität, Stuttgart. / Includes bibliographical references (p. [249]-275) and index.

Thesis (Ph. D.)--University of Washington, 2008. / Vita. Includes bibliographical references (p. 128-132).

Thesis (M.S.)--Rochester Institute of Technology, 2009. / Typescript. Accompanying CD-R contains the source code and related files for compilation and execution of the volume renderer developed in the thesis. Includes bibliographical references (leaves 110-112).

Thesis (Ph. D.)--Ohio State University, 2007. / Title from first page of PDF file. Includes bibliographical references (p. 229-239).

Développer des applications de streaming multimedia pour systèmes embarqués devient une tâche de plus en plus complexe. De nouveaux standards multimedia apparaissent régulièrement sur le marché pour supporter de meilleures résolutions et délivrer du contenu multimedia de meilleure qualité. Une conséquence est la complexification des plateformes matérielles et du développement logiciel. La méthode traditionnelle de débogage pour les applications de streaming multimedia est l'utilisation de traces d'exécution. Cependant, la quantité de données générée par les logiciels modernes augmente et les outils existants ne passent pas à l'échelle, ne permettent plus un débogage efficace. Dans cette thèse, nous nous focalisons sur de nouvelles techniques de visualisation enrichies par des résultats d'algorithmes de fouille de données afin de permettre une analyse efficace des traces d'exécution.Nous commençons par présenter les Slick Graphs, une technique de découpage et de lissage pour la visualisation de séries temporelles. Les Slick Graphs minimisent les artéfacts introduits par les techniques de lissage traditionnelles en utilisant le plus petit intervalle possible: les pixels. A travers une étude utilisateur, nous montrons que les Slick Graphs sont significativement plus rapides et plus précis avec des données périodiques. Nous proposons ensuite un nouveau système de visualisation interactive, TraceViz, pour explorer les traces d'exécution à différents niveaux de détails. Avec TraceViz, nous introduisons aussi un back-end permettant l'exploration interactive de trace d'exécution de taille importante. Nous fournissons une analyse de performance montrant que le back-end de TraceViz délivre des performances significativement meilleures que les back-end utilisés dans les outils de débogage disponibles aujourd'hui.Les traces contiennent aussi de nombreuses informations importantes qui peuvent être calculées avec des algorithmes de fouille de données comme par exemple l'existence de séquences d’événements répétées au cours de la trace ou des comportements périodiques. Cependant, même si les techniques de fouille de données permettent d'avoir une meilleure compréhension des traces d'exécution, leurs résultats sont difficiles à exploiter dû au grand nombre de motifs à examiner un par un manuellement. Nous proposons une nouvelle méthode d'analyse visuelle qui permet de visualiser les structures cachées dans une traces comme les séquences répétées et la périodicité d'un ensemble d’événements, permettant de rapidement avoir une compréhension fine de la trace. Enfin, nous montrons aussi comment notre méthode peut être appliquées à différents types de données, autres que les traces d'exécution. / Developing streaming multimedia applications on embedded systems becomes increasingly complex over time. New multimedia standards reach the market to support better resolutions and overall improved quality delivered to the end-user. Consequently, hardware platforms complexify and developing the software to fully exploit them becomes harder at each new generation. The traditional debugging method for streaming applications is the usage of execution traces. However, the amount of data generated by modern software largely increases and existing tools do not allow an efficient debugging process as they become unable to tackle large amounts of data. In this thesis, we focus on new interactive visualization techniques enriched by results of data mining algorithms for a more efficient analysis of execution traces for multimedia applications.First, we introduce Slick Graphs, a binning and smoothing technique for time series visualization. Slick Graphs mitigate the quantization artifacts, introduced by the traditional smoothing techniques, by using the smallest possible binning intervals, i.e. pixels. We compared Slick Graphs to traditional smoothing techniques in a user study and show that the Slick Graphs are significantly faster and more accurate when working with periodic data. We then propose a novel interaction visualization framework, TraceViz, to explore the execution traces at different level of details and integrate the Slick Graphs to provide a global overview of the trace. With TraceViz, we also introduce a fast back-end to support the interactive browsing of huge traces. We perform a performance analysis to show that the TraceViz back-end outperforms the back-end used in state-of-the-art debugging tools for execution traces.Execution traces contain meaningful information that can be computed using data mining techniques. A wide range of patterns can be computed and provide valuable information: for example existence of repeated sequences of events or periodic behaviors. However, while pattern mining approaches provide a deeper understanding of the traces, their results is hard to understand due to the large amount of patterns that have to be examined one by one. We propose a novel visual analytics method that allows to immediately visualize hidden structures such as repeated sets/sequences and periodicity, allowing to quickly gain a deep understanding of the trace. Finally, we also show how our method can be applied with different types of data than execution traces.

Técnicas de visualização de informações representam graficamente dados de um determinado domínio de aplicação e disponibilizam mecanismos para a interação com a representação gerada, a fim de que o usuário consiga interpretar e compreender as informações apresentadas. As técnicas de visualização multidimensionais, em particular, referem-se à visualização de informações onde cada elemento no conjunto de dados é descrito, ou caracterizado, por múltiplas variáveis (atributos), as quais devem ser codificadas em uma única estrutura visual. Embora existam diversas técnicas de visualização para representar dados multidimensionais, nenhuma delas apresenta o melhor desempenho para todos os tipos de tarefas. Uma estratégia interessante é analisar várias representações simultaneamente mantendo uma ligação semântica entre elas, de forma que ações realizadas sobre uma técnica se reflitam automaticamente nas demais. Não existe, atualmente, um sistema de múltiplas visualizações coordenadas específico para a representação de dados multidimensionais, que disponha de um amplo e variado conjunto de coordenações. Os sistemas que suportam a representação coordenada deste tipo de dados são de propósito geral e, devido a isto, oferecem poucas possibilidades de coordenação. O presente trabalho apresenta um estudo de técnicas de coordenação para sistemas de múltiplas visualizações, focando coordenação de representações visuais de dados multidimensionais. A partir do estudo dos diversos sistemas de visualização existentes, foi identificado um conjunto de coordenações que podem ser aplicadas entre visualizações multidimensionais. Esse trabalho apresenta estas coordenações e uma nova aplicação de múltiplas visualizações coordenadas (construída a partir do toolkit InfoVis), específica para a representação de dados multidimensionais, que implementa o conjunto de coordenações. A aplicação desenvolvida conta com um variado conjunto de coordenações e é altamente flexível, permitindo tanto a escolha das visualizações que representarão o conjunto de dados, bem como das coordenações ativas entre elas. Foram realizados alguns estudos de avaliação como ensaios de interação com usuários. Primeiramente, as implementações das técnicas de visualização suportadas pela aplicação desenvolvida foram avaliadas, com o objetivo de encontrar possíveis problemas de usabilidade. Grande parte dos problemas identificados nesta avaliação, principalmente os mais graves, foram solucionados logo após sua realização. Na seqüência, outro experimento, conduzido de maneira mais informal, avaliou algumas questões a respeito do uso das coordenações implementadas. Por fim, com um estudo de caso, verificou-se a aplicabilidade das visualizações multidimensionais suportadas pela aplicação para a exploração de dados de um domínio específico. / Information visualization techniques represent data of a specific domain graphically and provide mechanisms for interacting with this representation, allowing users to explore and understand their data. Multidimensional visualization techniques are employed to represent information where each data element is described by multiple variables (attributes) that must be mapped to a single visual structure. Although there are several techniques to display multidimensional data, none of them performs best for all kinds of tasks. An interesting strategy is to analyze multiple visualization techniques simultaneously with a semantic connection between them, so that actions performed on a representation are broadcasted to the others. This work investigates coordination techniques for multiple views systems, focusing the coordination of multidimensional data visualization. At the moment, there is not a specific coordinated-view system for multidimensional data representation that provides various forms of coordination. Systems allowing coordinated views are usually for general purpose. Due to this reason, they provide few coordination possibilities. Starting from the study of several visualization systems, we identified coordination forms that can be used between multidimensional visualizations. This work presents these coordinations and a new specific coordinated-view application (built with the InfoVis toolkit) for multidimensional data representation that implements the identified coordinations. Our application provides various coordination forms and is very flexible: it allows the user to choose the views that will display the dataset as well as the active coordinations. This work also presents some evaluation studies conducted as interaction tests with users. Firstly, the visualization techniques implemented in our application were evaluated in order to identify possible usability problems. Most of the problems identified in that study, mainly the more serious ones, were solved. Soon after, other experiment, conducted in a more informal way, evaluated some questions regarding the use of the implemented coordinations. Finally, we verified the suitability of the multidimensional visualizations supported by our application for a specific domain through a case study.

Em março de 2006 o Internet Research Task Force (IRTF) propôs uma metodologia para medições sobre tráfegos SNMP, a fim de identificar os padrões de uso desse protocolo. Contudo, essa metodologia apresenta algumas limitações, tais como: ausência de técnicas para visualização de dados, não especificação de formas de comparação dos resultados obtidos a partir de vários arquivos de tráfego e falta de integração entre as ferramentas necessárias para a execução da metodologia. Esta dissertação de mestrado apresenta uma proposta de arquitetura para uma ferramenta Web que automatiza, de forma integrada, a execução das etapas da metodologia do IRTF, buscando solucionar os problemas identificados nessa metodologia. Essa arquitura foi implementada em um software denominado Management Traffic Analyzer, e técnicas de visualização específicas para os resultados gerados a partir de análises de tráfegos SNMP foram desenvolvidas especialmente para essa ferramenta. Por fim, utilizou-se o Management Traffic Analyzer para o estudo de amostras de tráfego SNMP provenientes de duas grandes redes brasileiras: a do ponto de presença da Rede Nacional de Ensino e Pesquisa (RNP) no Rio Grande do Sul, e a da própria RNP. / In March 2006 the Internet Research Task Force (IRTF) proposed a methodology for measuring SNMP traffic traces. However, this methodology presents some limitations, such as: absence of data visualization techniques, non specification of forms for comparing results obtained from various traffic traces and lack of integration among the needed tools for the execution of the methodology. This masters dissertation proposes an architecture of a web tool which automatizes, in an integrated fashion, the execution of the IRTF methodology’s steps. This architecture was implemented in a software which was called Managament Traffic Analyzer, and visualization techniques specific for the results generated by the SNMP traffic traces analyses were developed specifically for this tool. Finally, the Management Traffic Analyzer was used for the study of SNMP traffic traces originated from two large brazilian networks: the one of the Brazilian National Education and Research Network (RNP) point of presence in the state of Rio Grande do Sul, and the one of the RNP itself.