Submitted by Filipe dos Santos (fsantos@pucsp.br) on 2016-09-22T18:33:26Z No. of bitstreams: 1 Marcelo Stoppa Augusto Corrêa.pdf: 12054164 bytes, checksum: 1d76d1664521ddad6653a226c9ccfe4b (MD5) / Made available in DSpace on 2016-09-22T18:33:26Z (GMT). No. of bitstreams: 1 Marcelo Stoppa Augusto Corrêa.pdf: 12054164 bytes, checksum: 1d76d1664521ddad6653a226c9ccfe4b (MD5) Previous issue date: 2016-04-15 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Pontifícia Universidade Católica de São Paulo / The modern spirit performed deep, structural transformations in society. The new and more fluid socio-political settings changed not only the way interpersonal relations happen, but also the arrangements by which individuals may attach to one another: groups, multitudes or communities have evolved to a new dimension. The socio-political changes and the growing sophistication of media render the groupings ephemeral, empowers the crowds and favors the creation of new and hybrid communicational and cultural codes through cultural identity clashes caused by faster economic exchange. The present work lays out a methodology that aims the collaboration to research relational, cultural and social phenomena in groups, to analyze crowd and community dynamics through an ensemble of techniques to crawl, analyze and visualize data and build relational and affinity topologies which we named affinimaps. This transdisciplinary methodology stands on Big Data, Open Data, analytics, ontologies and complex data visualization algorithms, as the technical axis; on the works of Jacob Levy Moreno and Timothy Leary, as the psychological and sociometric axis; and on infographics and topology, on the artistic axis. It intends to offer the representation of complex relations of different sorts of actors so as to transcend the vision and improve the detection of arrangement and behavior patterns. This way, it might contribute to the research conducted by different knowledge areas investigating the relationships between men and the world / O espírito moderno trouxe profundas transformações estruturais à sociedade. Os novos arranjos político-sociais, mais fluidos, mudaram não apenas as formas com que se dão as relações interpessoais, mas também como se formam os arranjos sociais pelos quais os indivíduos se vinculam uns aos outros: os grupos, as multidões e as comunidades não são os mesmos. As mutações sócio-políticas, com a sofisticação cada vez maior dos meios de comunicação, aumentam a efemeridade dos agrupamentos, dando às multidões o poder se auto-organizarem e as diferentes comunidades do planeta efetuam trocas econômicas com muita rapidez e que, pelo confronto de identidades culturais, tecem códigos de comunicação e cultura cada vez mais híbridos. O presente trabalho propõe a construção de uma metodologia que visa colaborar com a investigação de fenômenos relacionais, culturais e sociais nos grupos, bem como investigar a dinâmica nas multidões e nas comunidades por meio de um conjunto de técnicas de captura, análise e visualização de dados para a construção de topologias relacionais e de afinidade, que nomeamos afinimapas. Esta metodologia transdisciplinar apoia-se em Big Data, Open Data, analytics, ontologia e algoritmos de visualização de dados complexos, no eixo técnico; nas obras de Jacob Levy Moreno e Timothy Leary, no eixo psicológico e sociométrico; e na infografia e na topologia, no eixo artístico. Ela pretende fornecer a representação da complexidade das relações de diferentes tipos de atores para transcender a visão e favorecer a detecção de padrões de arranjos e comportamentos. Deste modo, deseja-se contribuir com as investigações conduzidas por diferentes áreas do saber que levem em conta as relações entre o homem e o mundo

Os rankings musicais destinam-se principalmente a fins de marketing, mas também ajudam os usuários a descobrir novas músicas, bem como a comparar artistas, álbuns, etc. Este trabalho apresenta uma ferramenta interativa para visualizar, encontrar e comparar rankings musicais usando diferentes técnicas além de exibir atributos das músicas. A técnica foi concebida após uma pesquisa remota que coletou dados sobre como as pessoas escolhem música. As técnicas de visualização tornam mais fácil obter informações sobre artistas e faixas, e também comparar os dados obtidos a partir dos dois principais rankings de música, Billboard e Spotify. A ferrament também permite a interação com dados pessoais. Resultados de experimentos conduzidos com usuários potenciais mostraram que a ferramenta foi considerada interessante, com um layout atrativo. Comparando com as formas tradicionais de visualizar rankings de músicas, usuários preferiram a ferramenta aqui desenvolvida, mas a diferença para Billboard e Spotify não foi grande. Entretanto, quando avaliada a usabilidade da ferramenta, os resultados foram melhores, principalmente no que se refere à filtragem e às técnicas de comparação. MusicVis foi também considerado fácil de aprender. / Music rankings are mainly aimed at marketing purposes but also help users in discovering new music as well as comparing songs, artists, albums, etc. This work presents an interactive way to visualize, find and compare music rankings using different techniques, including the display of music attributes. The technique was conceived after a remote survey we conducted to collect data about how people choose music. Our visualization makes easier to obtain information about artists and tracks, and also to compare the data gathered from the two major music rankings, namely Billboard and Spotify. The tool also provides interaction with personal data. The results obtained from experiments with potential users showed that the tool was considered interesting, with an attractive layout. Compared to traditional music ranking tools users preferred ours, but with not such a large difference from using Billboard or Spotify. However, when evaluating the usability of our tool, results are positive, mainly concerning to data filtering and comparison features. MusicVis was also considered easy to learn.

As pessoas estão ficando cada vez mais interessadas no uso de monitores de atividade. A quantidade de dados de indivíduos disponível está ajudando na expansão e desenvolvimento de novas aplicações e projetos de visualizações para ser usados em casa, em ciência (e.g. para entender melhor o comportamento de populações) ou em governos interessados em desenvolver cidades inteligentes. Nesse trabalho é apresentada uma visualização simples e intuitiva que permite a exploração de dados pessoais por pessoas comuns. Com foco em ajudar as pessoas a compreenderem a si mesmas melhor e perceber coisas novas sobre seus dados. A visualização construída neste projeto é baseada em metáforas de calendários, relógios e mapas, além de utilizar gráficos de barra para explorar dados crus. A exploração desses dados se dá pela interação entre essas visualizações. Para avaliar o produto do trabalho são apresentados dois casos de uso onde alguns usuários tiveram a oportunidade de observar e discutir suas informações de dois pontos de vista diferente: exploração de dados pessoais para auto-aperfeiçoamento e o uso do Everyday Visualization por cientistas da saúde. Em nenhum dos casos houve treinamento. As visualizações resultantes agregam diversas fontes de dados, indo além de outros trabalhos de visualização casual e pessoal. Os resultados promissores demonstram a viabilidade de tais técnicas para visualização de dados pessoais. / People are becoming increasingly more interested in the use of activity monitors and selfimprovement. The availability of individuals’ data is also pushing the development of new applications and data visualization projects to be used at home, in science (e.g. to better understand the behavior of populations) or for governments interested in developing intelligent cities. In this work, we present an easy and intuitive set of visualizations to allow the exploration of personal data by common people. We focus on helping people to know themselves better and to make sense of their own data. Our visualizations are based on the metaphors of calendars, clocks, and maps, as well as on the use of bar charts to explore raw data. Data exploration is therefore guaranteed by the interaction between them. In order to evaluate our work we present two use cases, where few users observe and discuss the data from different points of view: the exploration of personal data for self-improvement purposes, and the use of Everyday Visualization by health scientists. Both use cases were ran without any training session. The resulting visualization aggregates several different data sources, going beyond many of the personal and casual visualization works. The promising results achieved demonstrated the viability of the use of such techniques for personal data visualizations and sense making.

The overall goal of this dissertation is to elucidate the cellular and circuit mechanisms underlying flexible behavior in the prefrontal cortex. We are often faced with situations in which the appropriate behavior in one context is inappropriate in others. If these situations are familiar, we can perform the appropriate behavior without relearning how the context relates to the behavior — an important hallmark of intelligence. Neuroimaging and lesion studies have shown that this dynamic, flexible process of remapping context to behavior (task switching) is dependent on prefrontal cortex, but the precise contributions and interactions of prefrontal subdivisions are still unknown. This dissertation investigates two prefrontal areas that are thought to be involved in distinct, but complementary executive roles in task switching — the dorsolateral prefrontal cortex (dlPFC) and the anterior cingulate cortex (ACC). Using electrophysiological recordings from macaque monkeys, I show that synchronous network oscillations in the dlPFC provide a mechanism to flexibly coordinate context representations (rules) between groups of neurons during task switching. Then, I show that, wheras the ACC neurons can represent rules at the cellular level, they do not play a significant role in switching between contexts — rather they seem to be more related to errors and motivational drive. Finally, I develop a set of web-enabled interactive visualization tools designed to provide a multi-dimensional integrated view of electrophysiological datasets. Taken together, these results contribute to our understanding of task switching by investigating new mechanisms for coordination of neurons in prefrontal cortex, clarifying the roles of prefrontal subdivisions during task switching, and providing visualization tools that enhance exploration and understanding of large, complex and multi-scale electrophysiological data.

This body of work explores multiple technical and aesthetic methods of representing complex statistical information in an approachable visual language, bridging the boundaries between data science, graphic design and fine arts. Ordinary data charts are combined together with other charts and diagrams and transformed in unexpected ways in order to form the basic structure of mythical landscape scenes. Line plots over time become the rising and falling curves of hills and mountains, bar charts are morphed into industrial factories on the horizon, and bubble charts become billowing smoke, a forest of trees, or a school of fish. The hope is that the work will act as an engaging alternative to traditional data representation and will encourage curiosity and a fresh perspective.

Accounts of species’ range shifts in response to climate change, most often as latitudinal shifts towards the poles or upslope shifts to higher elevations, are rapidly accumulating. These range shifts are often attributed to species ‘tracking’ their thermal niches as temperatures in their native ranges increase. Our objective was to estimate the degree to which climate change-driven shifts in water temperature may increase the exposure of West Virginia’s native freshwater fishes to mountaintop removal surface coal mining. Mid-century shifts in habitat suitability for nine non-game West Virginia fishes were projected via Maximum Entropy species distribution modeling, using a combination of physical habitat, historical climate conditions, and future climate data. Modeling projections for a high-emissions scenario (Representative Concentration Pathway 8.5) predict that habitat suitability will increase in high elevation streams for eight of nine species, with marginal increases in habitat suitability ranging from 46-418%. We conclude that many West Virginia fishes will be at risk of increased exposure to mountaintop removal surface coal mining if climate change continues at a rapid pace.

Extreme Learning Machine (ELM) is a training algorithm for Single-Layer Feed-forward Neural Network (SLFN). The difference in theory of ELM from other training algorithms is in the existence of explicitly-given solution due to the immutability of initialed weights. In practice, ELMs achieve performance similar to that of other state-of-the-art training techniques, while taking much less time to train a model. Experiments show that the speedup of training ELM is up to the 5 orders of magnitude comparing to standard Error Back-propagation algorithm. ELM is a recently discovered technique that has proved its efficiency in classic regression and classification tasks, including multi-class cases. In this thesis, extensions of ELMs for non-typical for Artificial Neural Networks (ANNs) problems are presented. The first extension, described in the third chapter, allows to use ELMs to get probabilistic outputs for multi-class classification problems. The standard way of solving this type of problems is based 'majority vote' of classifier's raw outputs. This approach can rise issues if the penalty for misclassification is different for different classes. In this case, having probability outputs would be more useful. In the scope of this extension, two methods are proposed. Additionally, an alternative way of interpreting probabilistic outputs is proposed. ELM method prove useful for non-linear dimensionality reduction and visualization, based on repetitive re-training and re-evaluation of model. The forth chapter introduces adaptations of ELM-based visualization for classification and regression tasks. A set of experiments has been conducted to prove that these adaptations provide better visualization results that can then be used for perform classification or regression on previously unseen samples. Shape registration of 3D models with non-isometric distortion is an open problem in 3D Computer Graphics and Computational Geometry. The fifth chapter discusses a novel approach for solving this problem by introducing a similarity metric for spectral descriptors. Practically, this approach has been implemented in two methods. The first one utilizes Siamese Neural Network to embed original spectral descriptors into a lower dimensional metric space, for which the Euclidean distance provides a good measure of similarity. The second method uses Extreme Learning Machines to learn similarity metric directly for original spectral descriptors. Over a set of experiments, the consistency of the proposed approach for solving deformable registration problem has been proven.

Air pollution is a severe issue in many big cities due to population growth and the rapid development of the economy and industry. This leads to the proliferating need to monitor urban air quality to avoid personal exposure and to make savvy decisions on managing the environment. In the last decades, the Internet of Things (IoT) is increasingly being applied to environmental challenges, including air quality monitoring and visualization. In this thesis, we present CAVisAP, a context-aware system for outdoor air pollution visualization with IoT platforms. The system aims to provide context-aware visualization of three air pollutants such as nitrogen dioxide (NO2), ozone (O3) and particulate matter (PM2.5) in Melbourne, Australia and Skellefteå, Sweden. In addition to the primary context as location and time, CAVisAP takes into account users’ pollutant sensitivity levels and colour vision impairments to provide personalized pollution maps and pollution-based route planning. Experiments are conducted to validate the system and results are discussed.

Since the late 1800s, public data visualizations displaying election forecasts and results—such as the red and blue map of the United State—have presented an irreparably divided country. However, on November 8, 2016, the New York Times published a data visualization on their live presidential forecast page that broke over a century of visual expectations, inspiring many to tweet reactions to what popular media has dubbed the "jittery gauges." Not surprisingly, the tweets about this unique and difficult-to-interpret display were mostly negative. This paper argues, though, that the negative feedback indicates that the gauges, while imperfect, represent an important step away from visualizations that support the growing perception of party polarization. The key factor present in the gauges is the data design principle of uncertainty or possibility. If major news outlets were more thoughtful about introducing uncertain elements into visualizations of American politics, perhaps the nation could begin to imagine a political landscape that moves beyond red vs. blue, me vs. you.

Advanced Building Energy Data Visualization is a way to detect performance problems in commercialbuildings. By placing sensors in a building that collects data from example, air temperature and electricalpower, then makes it possible to calculate the data in Data Visualization software. This softwaregenerates visual diagrams so the building manager or building operator can see if for example thepower consumption is to high.A first step (before sensors are installed in a building) to see how the energy consumption is in abuilding can be to use a Benchmarking Tool. There is a number of Benchmarking Tools that is availablefor free on the Internet. Each tool have a bit different approach, but they all show how much energyconsumption there is in a building compared to other similar buildings.In this study a new web design for the benchmarking tool CalARCH has been developed. CalARCHis developed at the Berkeley Lab in Berkeley, California, USA. CalARCH uses data collected only frombuildings in California, and is only for comparing buildings in California with other similar buildingsin the state.Five different versions of the web site were made. Then a web survey was done to determine whichversion would be the best for CalARCH. The results showed that Version 5 and Version 3 was the best.Then a new version was made, based on these two versions. This study was made at the LawrenceBerkeley Laboratory.