Interactive Data Analysis and Model Exploration: A Visual Analytics Approach

• Activity Number: 678
• Type: Topic Contributed
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Remco Chang*+
• Companies: Tufts
• Keywords: Visual analytics ; information visualization ; interaction ; user modeling

Abstract

Visual analytics is a fast growing discipline that combines visualization, interaction, and data analysis in solving large and complex problems that require human understanding and analysis. Although each of these subfields has made significant advancements individually, approaches in visual analytics that combine them have demonstrated great benefits as well as revealing great challenges. In this talk, I will discuss methods that integrate interactive visualization and machine learning to assist the user in decision making, information retrieval, and knowledge discovery. Using similar principles, I will present projects relating to real-world problems in financial, biological, and other data analyses and show how data analysis can be combined with interaction and visual representation. Finally, I will highlight the importance of understanding a user's interaction in the context of reasoning and intent retrieval, and conclude with a presentation of its implications to mixed-initiative systems and extracting of a user's domain knowledge.