📝 SummarXiv

Abstract

Although recent efforts have developed accessible data visualization tools for blind and low-vision (BLV) users, most follow a "design for them" approach that creates an unintentional divide between sighted creators and BLV consumers. This unidirectional paradigm perpetuates a power dynamic where sighted creators produce non-visual content boundaries for BLV consumers to access. This paper proposes a bidirectional approach, "design for us," where both sighted and BLV collaborators can employ the same tool to create, interpret, and communicate data visualizations for each other. We introduce Py maidr, a Python package that seamlessly encodes multimodal (e.g., tactile, auditory, conversational) data representations into visual plots generated by Matplotlib and Seaborn. By simply importing the maidr package and invoking the maidr.show() method, users can generate accessible plots with minimal changes to their existing codebase regardless of their visual dis/abilities. Our technical case studies demonstrate how this tool is scalable and can be integrated into interactive computing (e.g., Jupyter Notebook, Google Colab), reproducible and literate programming (e.g., Quarto), and reactive dashboards (e.g., Shiny, Streamlit). Our performance benchmarks demonstrate that Py maidr introduces minimal and consistent overhead during the rendering and export of plots against Matplotlib and Seaborn baselines. This work significantly contributes to narrowing the accessibility gap in data visualization by providing a unified framework that fosters collaboration and communication between sighted and BLV individuals.