A new arXiv preprint, Interactive Mascot: A Scene-Centric Interaction Grammar for Data Visualizations, proposes a grammar for building interactions in visualization systems. The authors describe scene-centric systems that expose semantic components—such as marks, encodings, layouts, and axes—as first-class objects that users can directly manipulate.

According to the abstract, existing interaction abstractions are largely based on an approach that is not fully described in the available text. The implication is that prior methods do not fully leverage the scene-centric model. The paper's contribution appears to be a new grammar that makes such direct manipulation systematic.

This summary is based solely on the arXiv abstract, which is truncated. No details about the grammar's implementation, evaluation, or comparison to other systems are present in the provided source.