skills.rest
mutsumi-yamamoto
mutsumi-yamamotoCOMMUNITY
Explore all skills by this author →
GitHub

data-model

Perform predictive and inferential modeling with cross-validation, calibration, and SHAP explanations.

Updated Mar 3, 2026
Search Tags:#modeling#eda#machine-learning#causal-inference#hyperparameter-tuning#model-interpretability#timeseries
One-click install
npx skills add https://github.com/mutsumi-yamamoto/claude-data-analysis-marketplace --skill data-model-mutsumi-yamamoto
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: data-model
Source: https://github.com/mutsumi-yamamoto/claude-data-analysis-marketplace/tree/main/plugins/data-analysis/skills/data-model
Command: npx skills add https://github.com/mutsumi-yamamoto/claude-data-analysis-marketplace --skill data-model-mutsumi-yamamoto

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enables senior data scientists to execute Phase 6 modeling workflows that convert business hypotheses and context into validated predictive, time-series, clustering, and causal models and deliver reproducible evaluation artifacts and reports.

Core Features & Use Cases

  • Method selection guidance for regression, classification, time-series forecasting, clustering, and causal inference based on analysis goals and data characteristics.
  • Hands-on steps for descriptive and diagnostic statistics, class imbalance handling, baseline establishment, cross-validated model comparisons, hyperparameter optimization, probability calibration, and final test evaluation.
  • Time-series modeling with Prophet/SARIMA/Holt-Winters, unsupervised clustering and dimensionality reduction, SHAP-based model interpretability, and a mandatory execution log written to data/docs/06_modeling_report.md while updating analysis_context.md.
  • Use case: build and validate a demand-forecasting pipeline with automated model selection, calibration, SHAP explanations, and an executive-ready modeling report.

Quick Start

Run the data-model skill to perform Phase 6 modeling using the current analysis_context.md and project data and produce evaluation artifacts and data/docs/06_modeling_report.md.

Frequently Asked Questions about data-model

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

FAQPage Schema
How do I perform end-to-end predictive modeling and save reproducible reports?▼

End-to-end predictive modeling requires executing cross-validated model selection, hyperparameter optimization, and probability calibration. This Skill outputs validated predictive models, SHAP explanations, and a mandatory execution log saved as a reproducible report.

What's the best way to handle class imbalance and compare baselines in machine learning?▼

Handling class imbalance and comparing baselines involves applying resampling techniques during cross-validated model comparisons. This Skill establishes baseline performance, optimizes hyperparameters, and calibrates probabilities to ensure robust classification results.

Can I use Prophet or SARIMA for time-series forecasting within a CRISP-DM workflow?▼

Prophet, SARIMA, and Holt-Winters are supported for time-series forecasting within a CRISP-DM Phase 6 workflow. This Skill applies these algorithms to structured datasets to generate validated forecasts and reproducible evaluation artifacts.

How do I generate SHAP explanations for unsupervised clustering and causal inference tasks?▼

SHAP explanations apply to supervised models, while unsupervised clustering and causal inference require distinct diagnostic approaches. This Skill performs dimensionality reduction for clustering and observational causal inference, generating SHAP-based model interpretability for predictive outputs.

Does this modeling workflow require a specific environment setup or analysis context file?▼

This modeling workflow reads an existing analysis_context.md file and project data to execute properly. No external dependencies are required, but providing the context file ensures the modeling pipeline aligns with your business hypotheses and data characteristics.

Related Skills

Explore More Data & Analytics Skills→
mutsumi-yamamotomutsumi-yamamoto

data-define

Define business problems, structure issue trees, and design KPIs in analysis_context.md.

Community
Intermediate
sakamoto-family-smilesakamoto-family-smile

mle-workflow

Transform notebook experiments into production ML pipelines with data contracts and governance.

Community
Advanced
Phife726Phife726

ds-eda-process

Perform structured exploratory data analysis on CSV inputs using a CRISP-DM workflow.

Community
Advanced
zangxin75zangxin75

data-scientist

Perform statistical modeling, A/B testing, and time series forecasting.

Community
Advanced