Vega is a popular declarative language for creating interactive data visualizations. It supports reactive data transformations using its streaming dataflow architecture. Despite its wide-spread adoption, the exact semantics of Vega is subtle and poorly documented. This leads to difficult to understand error messages and incorrect or confusing visualizations.

This paper makes two contributions. First, we present a type system for the core data transformation language of Vega, which can prevent a range of common errors. Second, we define a graph-based operational semantics, providing a precise model of the streaming dataflow architecture of Vega. We show that our type system is sound with respect to the semantics.

Our work results in a practical checker for Vega specifications, but it also relates the dataflow architecture of Vega to well-studied ideas, including functional reactive programming and adaptive functional programming. The novelty of our work lies in considering functional reactive programming in the context of data visualization language and identifying what can be statically analysed in such environment.