Line graphs have been commonly used for visualizing temporal trends in time series data. Since comparing trends is one of the main tasks for analyzing multiple temporal trends, many efforts have been made to enhance visual representations of line graphs to help people efficiently compare multiple temporal trends. However, as the number of line graphs increases, the overlap makes it difficult to perform comparison and other analyses. DiffMatrix is a matrix-based interactive visualization designed to support effective analyses of a large number of time series data. It employs four visual representations for each cell in the matrix to show the difference between two time series—dual lines, diff line, diff area, diff heatmap—and a detail view to support more indepth analyses on individual line graphs. DiffMatrix allows users to seamlessly switch between these representations that best support their tasks.

When exploiting the power of node-link diagrams to represent real-world data such as web structures, airline routes, electrical, telecommunication and social networks, link congestion frequently arises. Such areas in the diagram—with dense, overlapping links—are not readable: connectivity, node shapes, labels, and contextual information are obscured. In response, graph-layout research has begun to consider the modification of link shapes with techniques such as link routing and bundling. We delve into the interactive techniques afforded by variant use of link curvature, delineating a six-dimensional design space that is populated by four families of interactive techniques: bundling, fanning, magnets, and legends. Our taxonomy encompasses existing techniques and reveals several novel link interactions. We describe the implementation of these techniques and illustrate their potential for exploring dense graphs with multiple types of links.

Exploring large network datasets, such as scientific collaboration networks, is challenging because they often contain a large number of nodes and edges in several types and with multiple attributes. Analyses of such networks are often long and complex, and may require several sessions by multiple users. Therefore, it is often difficult for users to recall their own exploration history or share it with others. GraphTrail is an interactive visualization for analyzing networks through exploration of node and edge aggregates that captures users’ interactions and integrates this history directly in the exploration workspace. To facilitate large network analysis, GraphTrail integrates aggregation with familiar charts, drag-and-drop interaction on a canvas, and a novel pivoting mechanism for transitioning between aggregates.

Tag clouds have proliferated over the web over the last decade. They provide a visual summary of a collection of texts by visually depicting the tag frequency by font size. In use, tag clouds can evolve as the associated data source changes over time. Interesting discussions around tag clouds often include a series of tag clouds and consider how they evolve over time. However, since tag clouds do not explicitly represent trends or support comparisons, the cognitive demands placed on the person for perceiving trends in multiple tag clouds are high. SparkClouds integrate sparklines into a tag cloud to convey trends between multiple tag clouds.

ManiWordle is a Wordle-based visualization tool that revamps interactions with the layout by supporting custom manipulations. ManiWordle allows people to manipulate typography, color, and composition not only for the layout as a whole, but also for the individual words, enabling them to have better control over the layout result.

iChase is a novel interactive visualization tool to provide administrators with better awareness of editing activities on Wikipedia. Unlike the currently used visualizations that provide only page-centric information, iChase visualizes the trend of activities for two entity types, articles and contributors. iChase is based on two heatmaps (one for each entity type) synchronized to one timeline. It allows users to interactively explore the history of changes by drilling down into specific articles and contributors, or time points to access the details of the changes.

We define interactive legends as a class of controls combining the visual representation of static legends and interaction mechanisms of widgets. As interactive legends start to appear in popular websites, we categorize their designs for common data types, and evaluate their performances compare to standard widgets.

In the conventional node-link tree visualization, the screen quickly becomes overcrowded as users open nodes that have too many child nodes to fit in one screen. To address this problem, we propose two extensions to the conventional node-link tree visualization: a list view with a scrollbar and a multi-column interface. We compared them against the conventional tree visualization interface in a user study.

FacetLens is a follow-on interface to prior work on FacetMap and PaperLens. FacetLens is an appropriate visualization tool for nearly any item/attribute based dataset expressed as a SQL database. Provided with a description of the types and locations of item metadata, FacetLens dynamically generates SQL queries to probe the data distributions and present them to the user in a data-rich, animated way. The user can  interactively click on values to apply filter criteria, and can use drag/drop operations to explore trends. FacetLens scales up to any size display by revealing progressively more detail as space grows.

FacetLens Video [Small size: .wmv (18.5MB), Large size: .wmv (43.3MB)]

FacetLens CHI2009 Presenation