Full text
| |
| Author(s): |
Pedro Henrique de Araújo Siqueira
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Computação |
| Defense date: | 2015-03-07 |
| Examining board members: |
Guilherme Pimentel Telles;
Ricardo da Silva Torres;
Maria Cristina Ferreira de Oliveira
|
| Advisor: | Guilherme Pimentel Telles |
| Abstract | |
Information Visualization is concerned with visually representing datasets. This representation should unveil relevant relations among data. Many applications can benefit from data visualization, for instance, searching for patterns in a dataset, finding a group of similar documents or searching for text containing certain subject. In this dissertation, we propose applying geographic metaphors to explain non-geographic data, taking advantage of the user's pre-existing ability with such metaphors and therefore easing the exploration of data. One of the most discussed metaphors in literature is landscape views that make use of the third dimension to map some data attribute, which naturally becomes more complex than a layout in two dimensions. We revisited landscape views in the light of progress made in visualization in recent years, as better multidimensional projection techniques that map the data as points in the plane with respect to similarity relations, and the ability to summarize datasets through word clouds and image mosaics. We have also developed two techniques for clustering and one space segmentation technique of projected points. We implemented two types of heat map, by density and by frequency of terms and we added the possibility of exploiting data hierarchically to landscapes, with the display of labels summarizing each group at each hierarchical level (AU) | |
| FAPESP's process: | 12/24634-6 - Visualization Using Geographic Metaphors |
| Grantee: | Pedro Henrique de Araújo Siqueira |
| Support Opportunities: | Scholarships in Brazil - Master |