Examensarbetet består av att skapa ett visualiseringsverktyg med syfte att bidra till en ökad förståelse hos en forskare att följa interaktion och beteende mellan multi-agenter när de traveserar ett träd. Av detta kommer forskaren enkelt kunna jämföra två studier och ta reda på om det är möjligt att minska overheadgapet mellan övre och undre gräns. Det främsta syftet med verktyget är det ska kunna visualisera specifika regler som forskaren har skrivit till en form av ett datavetenskapligt träd och låta en grupp agenter traversera igenom. Verktyget har skapats med hjälp av Java som programmeringsspråk samt JUNG visualisering-biblotek. Resultatet av detta arbete blev ett fungerande visualiseringsverktyg som följer forskarens regler samt har enkla traverseringsfunktioner. / The thesis consists of the creation of a visualization tool which purpose is to contribute to an increased understanding of following interactions and behaviorsbetween Multi-agents when traversing a tree. With this a scientist would be able to easily compare two studies to find out if it’s possible to shorten the overhead gap between the upper and lower limit. The prime purpose of the tool is to visualize specific rules written by the scientist, as a tree and letting a group of agents traverse in. The tool has been created in the programming language Java along with the visualization library JUNG. The result of this thesis was a functional visualization tool which follows the scientist’s rules as well as simple traversal functions.

Com base em métodos de extração de características de imagens e extração de vocabulários de textos, podemos empregar técnicas para posicionamento de dados multidimensionais no plano para mapear conjuntos de dados em espaços visuais, auxiliando usúarios na interpretação e análise dos dados. Alguns desses métodos constroem árvores de similaridade, impondo uma hierarquia sobre as relações entre as características extraídas dos dados. Em um ambiente de análise exploratória, é natural que se procurem métodos e técnicas capazes de manipular e interagir com os dados de forma rápida e eficiente. Nesse contexto, o trabalho visa implementar e aplicar técnicas conhecidas de navegação e interação em árvores no contexto de visualizações baseadas em posicionamento de pontos no plano. Em particular as técnicas NJ e MST, implementadas e utilizadas com sucesso na ferramenta PEx-Image, tornaram-se pontos chave para o auxílio na exploração dos dados através das apresentações radial e hiperbólica, implementadas juntamente com ferramentas de exploração. Este trabalho implementa e apresenta a capacidade exploratória dessas duas formas de apresentação de árvores sobre as visualizações NJ e MST. / Based on methods of feature extraction for images and vocabulary exploration for text, we can apply point placement techniques to multidimensional data in order to map data sets into visual spaces, assisting users on data analysis and interpretation. Some of these methods build similarity trees, imposing a hierarchy on the relationship between the characteristics extracted from data. In an exploratory analysis environment, it is natural to use methods and techniques capable of manipulating and interacting data quickly and eciently. In this context, this paper aims to implement and apply known techniques of tree navigation and interaction in the context of point placement visualizations. In particular the NJ and MST techniques, implemented and successfully used in the system PEx-Image, are the main focus for helping data exploration through Radial and Hyperbolic Layouts, implemented with exploration tools. This work implements Radial and Hyperbolic layouts to support exploration of NJ and MST views

Com base em métodos de extração de características de imagens e extração de vocabulários de textos, podemos empregar técnicas para posicionamento de dados multidimensionais no plano para mapear conjuntos de dados em espaços visuais, auxiliando usúarios na interpretação e análise dos dados. Alguns desses métodos constroem árvores de similaridade, impondo uma hierarquia sobre as relações entre as características extraídas dos dados. Em um ambiente de análise exploratória, é natural que se procurem métodos e técnicas capazes de manipular e interagir com os dados de forma rápida e eficiente. Nesse contexto, o trabalho visa implementar e aplicar técnicas conhecidas de navegação e interação em árvores no contexto de visualizações baseadas em posicionamento de pontos no plano. Em particular as técnicas NJ e MST, implementadas e utilizadas com sucesso na ferramenta PEx-Image, tornaram-se pontos chave para o auxílio na exploração dos dados através das apresentações radial e hiperbólica, implementadas juntamente com ferramentas de exploração. Este trabalho implementa e apresenta a capacidade exploratória dessas duas formas de apresentação de árvores sobre as visualizações NJ e MST. / Based on methods of feature extraction for images and vocabulary exploration for text, we can apply point placement techniques to multidimensional data in order to map data sets into visual spaces, assisting users on data analysis and interpretation. Some of these methods build similarity trees, imposing a hierarchy on the relationship between the characteristics extracted from data. In an exploratory analysis environment, it is natural to use methods and techniques capable of manipulating and interacting data quickly and eciently. In this context, this paper aims to implement and apply known techniques of tree navigation and interaction in the context of point placement visualizations. In particular the NJ and MST techniques, implemented and successfully used in the system PEx-Image, are the main focus for helping data exploration through Radial and Hyperbolic Layouts, implemented with exploration tools. This work implements Radial and Hyperbolic layouts to support exploration of NJ and MST views

<p>The Geologic <i>TimeScale Creator </i>(TSCreator) project has compiled a range of paleo-environmental and bio-diversity data which provides the opportunity to explore origination, speciation and extinction events. My PhD research has four major interconnected themes which include the visualization methods of evolutionary tree and the impacts of climate change on the evolution of life in longer and shorter timeframes: <b>(1) </b>Evolutionary range data of planktonic foraminifera and nannofossils over the Cenozoic era have been updated with our latest geological timescale. These evolutionary ranges can be visualized in the form of interactive, extensible evolutionary trees and can be compared with other geologic data columns. <b>(2) </b>A novel approach of integrating morphospecies and lineage trees is proposed to expand the scope of exploration of the evolutionary history of microfossils. It is now possible to visualize morphological changes and ancestor-descendant lineage relationships on TSCreator charts which helps mutual learning of these species based on genetic and bio-stratigraphic studies. <b>(3) </b>These evolutionary datasets have been used to analyze semi-periodic cycles in the past bio-diversity and characteristic rates of turnover. Well-known Milankovitch cycles have been found as the drivers of fluctuations in the speciation and extinction processes. <b>(4) </b>Within a shorter 2000-year time period, global cooling events might have been a factor of human civilization turnover. Using our regional and global cultural turnover time series data, the effect of climate change on human culture has been proposed. The enhancement of the evolutionary visualization system accomplished by this research will hopefully allow academic and non-academic users across the world to research and easily explore Earth history data through publicly available TSCreator program and websites. </p>