A fundamental problem in modern visual data analysis is how to build data exploration environments that support interactive exploration of large datasets. This problem has two opposing facets. From one side, the ever-growing complexity and size of datasets bring the need to provide complex navigation and visual summaries capabilities. On the other hand, human perception and cognition pose a challenge on how long the data handling and rendering loop can take. Unfortunately, the ability to produce compelling visual summaries, interaction mechanisms and interfaces has surpassed our capabilities to create techniques that support real-time data processing for visualization. As a result, there are limitations on the analysis that one can hope to perform interactively. In this talk, I will describe recent data handling strategies proposed to alleviate/overcome this challenge. The focus will be on novel indexing techniques that support real-time modeling, uncertainty quantification and event detection on large spatio-temporal datasets.
*Texto informado pelo autor.
Nivan Ferreira is a professor of computer science at Universidade Federal de Pernambuco. He received his PhD degree in Computer Science from New York University and held a postdoctoral scholar position at the University of Arizona. He previously attended the Universidade Federal de Pernambuco where he obtained his bachelor's degree in Computer Science and master's degree in Mathematics. His research interests are in the general areas of data visualization, visual analytics, big data and data science. His current research focuses on the interactive analysis of large spatio-temporal datasets and applications of data visualization for urban data analysis.