Diagrams-as-Code for Conceptual Modeling in Computational Problem Solving

We introduce an interface for a novel integration of diagrams-as-code and data analysis tools that links interactive conceptual modeling to data science tools for computational problem solving. The interface allows a user to seamlessly move between creating and updating their conceptual model of the problem while visualizing it, decomposing the problem with these modeling tools, and developing code to implement computational solutions and visualize results. All concepts identified in the models are visible throughout the process, allowing for easy navigation and model awareness. Implicit dependencies enable model checking and results aggregation. The integration with computational kernels offers an alternative to notebook-style interfaces that addresses several known shortcomings, while also providing polyglot integration among the computational components.