build-model

Automate CRISP-DM 4.3 model building with MLflow tracking and documentation.

Updated Mar 20, 2026

Search Tags:#reproducibility #experimentation #notebook #mlflow #model-building #data-preparation #crisp-dm

Ready to run

npx skills add https://github.com/thbraet/claude-template --skill build-model-thbraet

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: build-model
Source: https://github.com/thbraet/claude-template/tree/main/skills/build-model
Command: npx skills add https://github.com/thbraet/claude-template --skill build-model-thbraet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This CRISP-DM 4.3 workflow provides a repeatable process for constructing baseline and candidate models, with end-to-end experiment tracking and documentation.

Core Features & Use Cases

Baseline model construction, candidate model exploration, and experiment tracking with MLflow.
Generates a reproducible notebook and a formal model-building report detailing techniques, parameters, and outcomes.
Supports data pipelines, artifact management, and traceability from data preparation through evaluation.

Quick Start

Provide your prepared data and run the build-model workflow to generate the notebook and report.

Frequently Asked Questions about build-model

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build machine learning models with reproducible experimentation?▼

Model building in this CRISP-DM 4.3 workflow involves constructing baseline and candidate models with MLflow tracking. It automates experiment logging and generates a reproducible notebook alongside a formal model-building report.

What is the best way to track CRISP-DM model building experiments with MLflow?▼

Tracking CRISP-DM model building experiments with MLflow is handled automatically by this workflow. It logs model parameters, outcomes, and artifacts during candidate model exploration and outputs a summary document at docs/crisp-dm/4-modeling/.

Do I need prepared data before starting CRISP-DM 4.3 model building?▼

Yes, you need prepared data before starting CRISP-DM 4.3 model building. The workflow requires data pipeline outputs from prior phases to construct baseline models, generate the notebook at notebooks/4.3-model-building.ipynb, and track artifacts properly.

Can I generate a model building report and notebook together for data science projects?▼

Yes, you can generate a model building report and notebook together for data science projects. This workflow produces a reproducible notebook at notebooks/4.3-model-building.ipynb and a formal report at docs/crisp-dm/4-modeling/4.3-model-building.md simultaneously.

How does experiment tracking handle artifact management during model construction?▼

Experiment tracking handles artifact management by logging candidate models and data pipeline outputs directly to MLflow. This ensures proper traceability from data preparation through evaluation and maintains reproducibility across the generated artifacts.

What are the limitations of automating baseline model construction in a CRISP-DM workflow?▼

Automating baseline model construction in a CRISP-DM workflow is limited to projects requiring baseline and candidate models with MLflow logging. It focuses on the 4.3 modeling phase and does not handle data preparation or final deployment tasks.