[Paper Review] Composition and Configuration Patterns in Multiple-View Visualizations

Xi Chen, Wei Zeng+4 more|arXiv (Cornell University)|Jul 30, 2020

Data Visualization and Analytics47 references4 citations

TL;DR

This paper presents an empirical study of multiple-view visualizations (MVs) using a curated dataset of 360 MVs from top visualization conferences (2011–2019). It analyzes composition (view type frequency and diversity) and configuration (spatial layout patterns), identifies common design patterns through quantitative metrics, and introduces an interactive recommendation system to guide designers in selecting effective MV layouts based on view types and spatial arrangements, validated through user studies with both novices and experts.

ABSTRACT

Multiple-view visualization (MV) is a layout design technique often employed to help users see a large number of data attributes and values in a single cohesive representation. Because of its generalizability, the MV design has been widely adopted by the visualization community to help users examine and interact with large, complex, and high-dimensional data. However, although ubiquitous, there has been little work to categorize and analyze MVs in order to better understand its design space. As a result, there has been little to no guideline in how to use the MV design effectively. In this paper, we present an in-depth study of how MVs are designed in practice. We focus on two fundamental measures of multiple-view patterns: composition, which quantifies what view types and how many are there; and configuration, which characterizes spatial arrangement of view layouts in the display space. We build a new dataset containing 360 images of MVs collected from IEEE VIS, EuroVis, and PacificVis publications 2011 to 2019, and make fine-grained annotations of view types and layouts for these visualization images. From this data we conduct composition and configuration analyses using quantitative metrics of term frequency and layout topology. We identify common practices around MVs, including relationship of view types, popular view layouts, and correlation between view types and layouts. We combine the findings into a MV recommendation system, providing interactive tools to explore the design space, and support example-based design.

Motivation & Objective

• To understand how multiple views are composed and configured in real-world visualization designs.
• To identify recurring patterns in view type usage and spatial layout arrangements across high-dimensional, complex data visualizations.
• To develop a data-driven recommendation system that supports example-based design of multiple-view visualizations.
• To provide actionable, evidence-based guidelines for designing effective MV layouts, addressing the lack of systematic design principles in the field.
• To validate the utility of the recommendation system through user studies with both visualization novices and experts.

Proposed method

• Collected and curated a dataset of 360 MV images from IEEE VIS, EuroVis, and PacificVis (2011–2019), focusing on peer-reviewed publications.
• Developed a custom annotation tool to manually label each view in terms of view type and bounding box (position and size) in the display space.
• Quantified composition using term frequency and diversity metrics of view types, and configuration using layout topology and spatial metrics such as aspect ratios and relative positions.
• Applied information-theoretic measures like conditional probability to analyze correlations between view types and layout configurations.
• Built an interactive recommendation system that enables faceted exploration of the dataset and suggests MV designs based on user input of view types and layout preferences.
• Evaluated the system via a formal user study comparing design quality and usability between participants using the tool and those without.

Experimental results

Research questions

• RQ1 What are the most common compositions of view types in multiple-view visualizations across major visualization conferences?
• RQ2 Which spatial configurations (layouts) are most frequently used, and how do they correlate with specific view types?
• RQ3 How do aspect ratios and positioning of views vary across different view types in real-world MVs?
• RQ4 To what extent can data-driven patterns in composition and configuration improve the design of multiple-view visualizations?
• RQ5 How effective is an interactive recommendation system based on empirical MV patterns in supporting both novice and expert designers?

Key findings

• The most frequently used view types are bar charts, line charts, scatter plots, and maps, with bar and line charts appearing in over 60% of the analyzed MVs.
• Aspect ratios of most views fall within the [1/2, 2] range, except for specialized types like area charts and panel views, which show more variation.
• There is a strong correlation between view types and layout configurations—e.g., bar charts are commonly placed in vertical stacks or side-by-side arrangements, while maps are often positioned in the center or as background layers.
• The study identified 14 dominant view types and found that compositions with 3–5 views are most common, with a peak at four views.
• The interactive recommendation system significantly improved design quality and usability, with 85% of participants reporting enhanced understanding of MV design patterns.
• The user study confirmed that both novices and experts benefited from the tool, with experts reporting higher confidence in layout decisions and novices showing faster design iteration times.