r-model-builder

What problem does it solve?

Reusable, battle-tested R modeling patterns that let teams build advanced model-helper layers (specification, formula/builders, data-slicing, fit wrappers, diagnostics, and post-estimation adapters) in a consistent, testable way across Bayesian, causal-design, meta-analysis, lavaan SEM, and tidymodels workflows.

Core Features & Use Cases

• Modular spec and formula builders that separate estimand, covariates, and random effects for reliable modeling across diverse families (brms, lavaan, metafor, marginaleffects, and tidymodels).
• Explicit data-materialization boundaries that collect only model-ready frames from large datasets or Arrow Parquet sources, with a clear path from specification to final reporting.
• A layered architecture that keeps fit, diagnostics, post-estimation, and deployment adapters as independent surfaces, enabling robust testing and reusable components.

Quick Start

Define a modeling spec, build the right-hand side from validated terms, prepare a ready-to-fit data frame, fit the model with data =, and surface tidy post-estimation outputs.