DEFOG: A System for Data-Backed Visual Composition

As users analyze data, visualization is important for both generating insight during exploration and displaying information for presentation purposes. While visualization systems have been very successful at the latter, their full potential as an aid for data exploration has yet to be realized. Existing systems have often focused on techniques for displaying information, requiring data to conform to specific file formats or allowing very limited manipulation of the data. In addition, most systems that provide a stronger link between data and visualization sacrifice the freedom of arbitrary composition. In this paper, we present DEFOG, a system that aims to more tightly integrate data exploration and visualization. DEFOG was designed to manipulate objects and as such, it is able to handle a wide range of data types. In addition, it allows flexible data manipulation through both visual and programmatic operations, all integrated in the same environment. By allowing both objects and their relationships to be displayed and combined, DEFOG provides great freedom for visual composition. To that end, it combines information visualization with infrastructure to support a variety of data including in-memory objects and operations found in vector-based graphics editors like select, move, scale, group, and copy-and-paste. We show that this system provides a natural way to use and combine data visualization techniques to explore a variety of data types, and present case studies that demonstrate its use in analyzing collections of workflow specifications and biochemical pathways. DEFOG: A System for Data-Backed Visual Composition Lauro Lins David Koop Juliana Freire Cláudio Silva Abstract— As users analyze data, visualization is important for both generating insight during exploration and displaying information for presentation purposes. While visualization systems have been very successful at the latter, their full potential as an aid for data exploration has yet to be realized. Existing systems have often focused on techniques for displaying information, requiring data to conform to specific file formats or allowing very limited manipulation of the data. In addition, most systems that provide a stronger link between data and visualization sacrifice the freedom of arbitrary composition. In this paper, we present DEFOG, a system that aims to more tightly integrate data exploration and visualization. DEFOG was designed to manipulate objects and as such, it is able to handle a wide range of data types. In addition, it allows flexible data manipulation through both visual and programmatic operations, all integrated in the same environment. By allowing both objects and their relationships to be displayed and combined, DEFOG provides great freedom for visual composition. To that end, it combines information visualization with infrastructure to support a variety of data including in-memory objects and operations found in vector-based graphics editors like select, move, scale, group, and copyand-paste. We show that this system provides a natural way to use and combine data visualization techniques to explore a variety of data types, and present case studies that demonstrate its use in analyzing collections of workflow specifications and biochemical pathways. As users analyze data, visualization is important for both generating insight during exploration and displaying information for presentation purposes. While visualization systems have been very successful at the latter, their full potential as an aid for data exploration has yet to be realized. Existing systems have often focused on techniques for displaying information, requiring data to conform to specific file formats or allowing very limited manipulation of the data. In addition, most systems that provide a stronger link between data and visualization sacrifice the freedom of arbitrary composition. In this paper, we present DEFOG, a system that aims to more tightly integrate data exploration and visualization. DEFOG was designed to manipulate objects and as such, it is able to handle a wide range of data types. In addition, it allows flexible data manipulation through both visual and programmatic operations, all integrated in the same environment. By allowing both objects and their relationships to be displayed and combined, DEFOG provides great freedom for visual composition. To that end, it combines information visualization with infrastructure to support a variety of data including in-memory objects and operations found in vector-based graphics editors like select, move, scale, group, and copyand-paste. We show that this system provides a natural way to use and combine data visualization techniques to explore a variety of data types, and present case studies that demonstrate its use in analyzing collections of workflow specifications and biochemical pathways.