ATWL: A Formal Language for Representing, Comparing, and Reusing Visual Analytics Workflows
Researchers have developed a new formal language called ATWL (Artifact-Transform Workflow Language) to represent visual analytics workflows. This language aims to standardize the description of complex processes like data transformation and human interpretation, moving beyond unstructured prose. ATWL utilizes a modular ontology of artifact types and transformation intents, enabling systematic comparison, reuse of strategies, and better training for novices. The researchers demonstrated its utility by extracting workflows from research papers using LLM agents and showed that ATWL representations provide more explicit structure and compactness compared to traditional paper formats when used with LLMs. AI
IMPACT Standardizes complex visual analytics processes, potentially improving LLM-driven research and analysis.