The ladder of abstraction in statistical graphics

2025-01-12

AI Summary

Statistical graphics often suffer from poor interpretability due to opaque abstraction levels. To address this, we propose the “Ladder of Abstraction” design paradigm—an original, systematic adaptation of linguistic abstraction theory to statistical visualization. Our method enables progressive graphic construction: starting from concrete, instance-level plots (e.g., a single-point scatterplot) and iteratively embedding them within increasingly general frameworks (e.g., distributions, trends, statistical models). Integrating equation-based modeling visualization with real-world social data analysis, the paradigm supports case-driven, hierarchical graphic development. Empirical evaluations—including complex trajectory model visualization and income–voting relationship analysis—demonstrate substantial improvements in audience comprehension efficiency and acceptance. This work establishes a principled, actionable, and transferable theoretical and practical framework for layered statistical graphic design.

Technology Category

• Reasoning under Uncertainty: Graphical Models
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Statistical Relational/Logic Learning

Application Category

• Web Mining and Content Analysis: Web data visualization
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Graphical forms such as scatterplots, line plots, and histograms are so familiar that it can be easy to forget how abstract they are. As a result, we often produce graphs that are difficult to follow. We propose a strategy for graphical communication by climbing a ladder of abstraction (a term from linguistics that we borrow from Hayakawa, 1939), starting with simple plots of special cases and then at each step embedding a graph into a more general framework. We demonstrate with two examples, first graphing a set of equations related to a modeled trajectory and then graphing data from an analysis of income and voting.

Problem

• Addressing abstraction in statistical graphics communication
• Proposing ladder strategy for clearer graphical representation
• Improving interpretability of scatterplots and line graphs

Innovation

• Climbing ladder of abstraction
• Embedding graphs into general framework
• Demonstrating with equation and data examples