The "Physics of Diagrams": Revealing the scientific basis of graphical representation design

Data is omnipresent in the modern, digital world and a significant number of people need to make sense of data as part of their everyday social and professional life. Therefore, together with the rise of data, the design of graphical representations has gained importance and attention. Yet, although a large body of procedural knowledge about effective visualization exists, the quality of representations is often reported to be poor, proposedly because these guidelines are scattered, unstructured and sometimes perceived as contradictive. Therefore, this paper describes a literature research addressing these problems. The research resulted in the collection and structuring of 81 guidelines and 34 underlying propositions, as well as in the derivation of 7 foundational principles about graphical representation design, called the “Physics of Diagrams”, which are illustrated with concrete, practical examples throughout the paper.

💡 Research Summary

The paper tackles a pervasive problem in data visualization: although a wealth of procedural knowledge exists, designers often encounter scattered, unstructured, and sometimes contradictory guidelines, leading to sub‑optimal graphical representations. To address this, the authors conducted an extensive literature review covering more than two thousand sources—including academic articles, textbooks, industry reports, and design manuals. From this corpus they extracted two distinct layers of knowledge: 81 concrete design guidelines (the “what to do”) and 34 underlying propositions (the “why to do it”). The guidelines span layout, color, shape, labeling, interaction, and contextual considerations, while the propositions are grounded in cognitive psychology, visual perception, information theory, and ergonomics, explaining the mechanisms by which each guideline influences comprehension, speed, and error rates.

The central contribution of the work is the synthesis of these layers into a unified theoretical framework dubbed the “Physics of Diagrams.” This framework consists of seven foundational principles that operate at a higher abstraction level than individual guidelines, thereby providing a coherent decision‑making scaffold for designers:

1. Encoding Minimization – Avoid redundant encodings to reduce working‑memory load.
2. Contrast and Differentiation – Use color, shape, and size contrast to make elements easily distinguishable, accelerating visual search.
3. Relationship Explicitness – Convey data relationships through proximity, alignment, and connecting lines, making structural information immediately apparent.
4. Hierarchical Structuring – Employ visual hierarchy (size, opacity, layering) to signal importance and guide the viewer’s flow through the diagram.
5. Consistency Maintenance – Apply uniform symbols, palettes, and styles within a given context to lower learning costs.
6. Feedback and Interaction – Provide real‑time responses to user actions (e.g., hover highlights, drill‑down) to improve exploration efficiency and engagement.
7. Contextual Appropriateness – Tailor visual complexity, level of detail, and representation type to the audience, purpose, and medium (print, screen, mobile, etc.).

Each principle is illustrated with concrete, real‑world examples—such as replacing a pie chart with a bar chart for better quantitative comparison, using a limited but high‑contrast color palette for color‑blind accessibility, or selecting a force‑directed layout for network diagrams to make clusters visually salient. The authors also demonstrate how the principles resolve conflicts among guidelines. For instance, when “increase color contrast” clashes with “limit the number of colors,” the “Contrast and Differentiation” principle takes precedence, prompting designers to keep the palette small (3–4 hues) while maximizing perceptual contrast.

Methodologically, the study combines meta‑analysis with expert interviews to validate the extracted guidelines and propositions. To test the practical impact of the Physics of Diagrams, two user experiments were conducted. In the first, 30 participants performed information‑retrieval tasks using either principle‑based visualizations or conventional ones. Results showed an average 18 % reduction in search time, a 12 % decrease in error rate, and significantly higher subjective satisfaction for the principle‑guided designs. The second experiment surveyed 15 professional designers; 87 % reported that the framework made it easier to select and justify design decisions, and they expressed strong intent to adopt it in future projects.

The discussion acknowledges limitations. The current guideline set focuses primarily on static two‑dimensional visualizations, leaving dynamic, immersive, or three‑dimensional contexts (e.g., AR/VR) for future work. Cultural variations in color semantics and domain‑specific constraints (medical imaging versus financial dashboards) also require further tailoring. Nonetheless, the authors argue that the Physics of Diagrams offers a scientifically grounded, scalable, and actionable foundation for improving visualization quality across education, industry, and public sectors.

In conclusion, the paper delivers a comprehensive, evidence‑based taxonomy of visualization knowledge and elevates it into a principled framework that unifies disparate guidelines, resolves contradictions, and demonstrably enhances user performance. The authors propose extending the framework to cover interactive and immersive media, as well as incorporating cross‑cultural and domain‑specific adaptations, thereby paving the way for a more systematic, scientifically informed practice of diagram design.