communication mediumLeipzig Symposium on Visualization in Applications (LEVIA) 2020
The need to use data visualization and visual analysis in various ﬁelds has led to the development of feature rich standalone applications such as Tableau and MS Power BI. These applications provide ready-to-use functionality for loading, analyzing and visualizing data, even for users who are not familiar with programming and scripting. Meanwhile, data scientists have to combine many different tools and techniques in their daily work, since no standalone application can yet cover the entire workﬂow. As a result, a rich landscape of open source libraries is available today, covering various tasks from data analysis to modeling and visualization. To combine the best of two worlds, interfaces for scripting languages have been integrated into standalone applications in recent years. We analyzed which interfaces to six common scripting languages are offered. The interfaces offer different levels of integration and therefore support different steps of the data science workﬂow. In this paper we investigated the integration levels of script languages in standalone applications and divided them into four groups. We used this classiﬁcation to evaluate 13 standalone visual analysis applications currently available on the market. We then analyzed which groups of applications best support which steps in the data science workﬂow. We found that a tight integration of scripting languages can especially support the explorative analysis and modeling phase of the data science workﬂow. We also discuss our results in the light of visual analysis research and give suggestions for future research directions.
visual analytics applications, scripting language interfaces