Due to the growth and complexity of data, the analysis process become even hard. Visualization has an important role in this scenario, in which visual metaphors are employed to aid in the knowledge discovering. For multidimensional data, one of the most used techniques are the multidimensional projections, that map data from high dimension to data in low dimension aiming at preserving similarity and neighborhood relations present in the multidimensional space. Due to the dimensionality reduction, overlaps among the graphical representations arise and the exploration process is impaired. Although the literature presents many overlap removal techniques, the space required to generate layouts without overlap is proportional to the size of the dataset. In order to minimize such limitation, this project aims at elaborating a new technique for multilevel exploration on visual representations generated by multidimensional projection techniques.
News published in Agência FAPESP Newsletter about the scholarship: