Towards accessible chart visualizations for the non-visuals: Research, applications and gaps

Mandhatya Singh, Muhammad Suhaib Kanroo, Hadia Showkat Kawoosa, Puneet Goyal

Abstract: Chart Visualizations (CharVis) such as charts/plots and diagrams are commonly used in documents for representing the underlying quantitative information. However, the inaccessibility of such visualizations exemplify one of the rife challenges of information access for Blind and Visually Impaired People (BVIP). The existing BVIP-related assistive technologies (ATs) are capable enough to provide the accessibility of textual components; however, for CharVis, it is of concern. Unlike textual components, CharVis comprise critical compressed data and requires perspicacious reverse-engineering schemes to output the raw data table used initially for creation. An intelligent and automated BVIP-compatible CharVis understanding scheme requires extraction of raw underlying data and presenting it into BVIP-compatible representation, i.e., summarized audio form. With the recent advancements in rapidly-growing AI-domain, several frameworks have been proposed in the literature for accurate extraction of the raw content from CharVis. Most of the existing related work and surveys emphasize only the visualization-related aspects without considering the apprehensions regarding inclusivity of these visualization schemes for BVIP. This survey aims to analyze various research methodologies on the CharVis understanding process, and related existing and potential assistive applications in a threefold outcome manner - (1) Research: We provide a perspicuous investigation of state-of-the-art research methodologies for CharVis understanding. (2) Applications: We provide a detailed rubric analysis of various applications and compatible assistive solutions (3) Gap: We summarize the challenges and the gaps between the research and application domain and provides new insights/pointers for future research. Additionally, the survey presents a consolidated list of datasets, software/apps, and hardware sources for supporting future research.