Visual Analytics
Visual analytics is the science of analytical reasoning supported by interactive visual interfaces, coupling human judgment with automated data analysis to make sense of large, complex data.
Definition
Visual analytics is the combination of automated analysis techniques with interactive visualizations to support analytical reasoning, enabling people to explore, understand, and draw conclusions from large and complex datasets through a human-in-the-loop process.
Scope
This topic covers the integration of visualization with computation for analysis: the visual analytics process that interleaves automated methods and interactive visual exploration, the support of sensemaking and hypothesis generation, and the handling of large, heterogeneous, and uncertain data. It does not cover the static encoding or interaction techniques in isolation, treated under their own topics, nor the machine-learning algorithms themselves, which belong to artificial intelligence.
Core questions
- How does visual analytics combine human reasoning with automated analysis?
- What is the sensemaking process and how can tools support it?
- How is the visual analytics process structured as an iterative loop?
- How are scale, heterogeneity, and uncertainty in data handled?
Key concepts
- analytical reasoning
- sensemaking
- visual analytics process
- human-in-the-loop analysis
- information foraging
- automated analysis integration
- hypothesis generation
- uncertainty in analysis
Key theories
- The visual analytics agenda and process
- Visual analytics was framed as detecting the expected and discovering the unexpected by tightly coupling interactive visualization with automated analysis; its process iterates between models, visualizations, and human interaction to refine insight.
- Sensemaking and the analytic loop
- Pirolli and Card modeled analysts' sensemaking as cycles of foraging for information and building and testing hypotheses, identifying leverage points where visual and computational tools can most help.
- Human in the loop with automation
- Rather than fully automating analysis, visual analytics keeps the human in the loop, using computation to summarize and surface patterns while relying on human judgment for interpretation, context, and decisions.
Clinical relevance
Visual analytics tools help analysts in domains such as intelligence, cybersecurity, public health, and business intelligence make sense of data too large or complex for either visualization or automation alone, supporting decisions by combining computational power with human insight.
History
Visual analytics was named and defined in the 2005 research agenda Illuminating the Path, motivated by the need to analyze massive, heterogeneous data. Keim and colleagues refined its definition and process, and sensemaking models such as Pirolli and Card's grounded it in cognitive theory, establishing visual analytics as a field bridging visualization, analysis, and human reasoning.
Key figures
- James J. Thomas
- Kristin A. Cook
- Daniel A. Keim
- Peter Pirolli
- Stuart K. Card
Related topics
Seminal works
- thomas2005
- keim2008
- pirolli2005
Frequently asked questions
- How is visual analytics different from information visualization?
- Information visualization focuses on representing data visually for human perception. Visual analytics is broader, integrating those visualizations with automated analysis methods, such as statistics and machine learning, in an interactive loop, so that human reasoning and computation work together on large, complex problems.
- Why keep a human in the loop instead of fully automating analysis?
- Automated methods are powerful at finding patterns in large data but lack context, judgment, and the ability to ask the right questions. Visual analytics keeps people in control of interpretation and decisions while using computation to handle scale, combining the strengths of both.