Description
Making data visualizations accessible to blind and low-vision (BLV) users has been an active area of research for many years. Recent advances in multimodal large language models (MLLMs), particularly vision-language models, offer promising new opportunities for improving accessibility. While prior work has primarily focused on making individual charts accessible, there remains a significant gap in supporting complex, multi-view visualizations such as interactive dashboards. How can dashboards be made accessible to the BLV users? In this project, we explore how can we leverage AI in the dashboard exploration process for BLV users?
Tasks
Create a visualization dashboard that is powered by MLLMs and can explain the visualizations and their interactions to the BLV users in a way that makes sense to them. The created prototype needs to be evaluated as well.
Requirements
- Knowledge of English language (source code comments and final report should be in English)
- Knowledge of web based technologies and visualizations is necessary
- Knowledge of Multimodal Large Language Models (MLLMs) is required
Environment
The project should be implemented as a standalone application, web-based application.