Better Information Visualization Software Through Packages for Data Science Ecosystems

Abstract

Good software development practices are important factors for the successful translation of visualization research into software. This paper argues for the creation of packages for data science ecosystems, with Python and R as case studies, as a way to employ existing tools and infrastructure towards better information visualization software. The paper describes open practices, sustainability and FAIR software to motivate package development. The ecosystems of Python and R are then reviewed based on general software development aspects and how common features of visualization software, such as rendering and interactivity, are supported. It concludes with the software engineering benefits related to creating packages in Python and R and initiatives to overcome obstacles that may hinder the development of better software.