Towards Visualisation Specifications from Multilingual Natural Language Queries using Large Language Models

Abstract

In this paper, we present an empirical demonstration of a prompt-based learning approach, which utilizes pre-trained Large Language Models to generate visualization specifications from user queries expressed in natural language. We showcase the approach's flexibility in generating valid specifications in languages other than English (e.g., Spanish) despite lacking access to any training samples. Our findings represent the first steps towards the development of multilingual interfaces for data visualization that transcend English-centric systems, making them more accessible to a wider range of users.