Graphics and Visualization (discontinued) - overview
Visualization is the process of representing data graphically and interacting with these representations in order to gain insight into the data. Traditionally, computer graphics has provided a powerful mechanism for creating, manipulating, and interacting with these representations. At IBM, graphics and visualization research addresses the problem of converting data into compelling, revealing, and interactive graphics that suit users' needs.
Our research includes studying visual analytics, developing languages and models of interaction with visualizations, designing novel information visualizations for smarter visual analytics, developing new representations of 3D geometry, strategies for collaborative and social visualization, and designing software systems that support a full range of display formats ranging from smartphones to immersive multi-display visualization environments.
Projects in Highlight
Many Eyes is a public web site that allows users to gather data, visualize it, and discuss their visualizations. We use the site as an experimental platform to test our hypotheses about the ability of visualizations to spur communication and social interaction, and how that activity may yield new insights into data.
IBM Interactive Maps Technology provides scalable interactive visualization of geo-located big data. The technology enables to build out-of-the-box visualizations as well as customized interactive visualization for integrated solutions. IBM interactive Maps Technology integrates with IBM Big Data Platform to support visualization of static GIS data, real-time streaming data, historical data, and data in motion .
Leveraging psycho-linguistic research, we now have the ability to automatically derive various personal intrinsic traits, including one’s personality, fundamental needs, basic values, and emotional status, from their linguistic footprints left on social media and social networking sites. In this project, we present a novel visual text analysis system, PETALS (People Essential Traits Analysis and Learning System), which helps users understand and verify various derived intrinsic traits of individuals. First, it provides an interactive visual analytic tool that helps users understand and explore various types of intrinsic traits of individuals derived from social media. Second, it uses a novel visual metaphor that encodes an individual’s psychological portrait, including multi-dimensional values of personal traits, and various meta properties of the derived traits (e.g., accuracy and transience). Third, the tool allows users to examine linguistic evidence at multiple levels, which helps the users comprehend how the traits are derived and assess the quality of the derived traits.
Big Picture project leverages advanced data summarization algorithms inspired by the latest advances in cognitive science and machine intelligence. The algorithms create semantically meaningful abstraction of raw data called Conceptual Spaces (CS), where regions of the spaces correspond to high level semantic properties. Visualizing data at a higher level of abstraction not only reduces visual clutters, but also bring the visualization closer to users' tasks. Big Picture also employee natural visual metaphors for the problem domain, resulting in attractive and consumable interactive visualization.
Many Bills is a web based visualization that aims to make congressional legislation easier to digest. It presents bills from the House and Senate organized into collections and split into sections which are color coded and labelled to indicate what topic each section is about. It also provides a set of features designed to make it easier to find interesting or unusual parts of bills and communicate your findings to others.
People make a slew of large and small decisions that significantly impact their business and quality of life. For example, a hotel manager might analyze hotel transactions and customer feedback to decide on the hotel’s future promotion programs. Likewise, a potential customer may analyze and compare features and prices of similar products before making a purchase. Often specialized skills and elaborate infrastructure are required to examine data and derive meaningful and actionable insights, this however limits the ability of average users to take advantage of today’s analysis tools. The goal of n.Mind is visual analytics to support decision making for the masses. We are developing an intelligent collaborative visual analytic system called n.Mind, which enables a recipe-driven, interactive visual investigation. In this approach, users start with their analytic goals/tasks at a high level and focus on examining the analysis results, while the system handles the details from extracting the data, choosing the method of analysis, to creation of the right visualization. We are conducting research in several aspect of this, from modeling of users visual analytic tasks, to enabling collaborative use of analytic recipes, and devising a system with a strong foundation on a declarative visual analytics language.
Electronic medical records contain a wealth of information that can help drive both healthcare policy decisions as well as individualized treatment plans. We are exploring various tools to help users analyze and visualize aggregated patient data. Our interative visual analysis tools are designed to help clinicians and medical administrators gain insights from historical data to better understand what treatments yield the best results in terms of both efficacy and resource efficiency. We have designed several interactive visualization prototypes to address different challenges within this domain, such as the Outflow visualization shown here which supports individualized temporal outcome analysis.
SaNDVis is a novel visual social network analytics tool that supports expertise location, team building, and team coordination in the enterprise. The interface integrates social position, evidence, and facets which allows users to reflect on existing relationships as well as build new relationships in an enterprise setting.