inno-experiment-analysis

Analyze ML experiment logs and generate statistical results and figures.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/t2ance/dr-claw-plugin --skill inno-experiment-analysis-t2ance
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: inno-experiment-analysis
Source: https://github.com/t2ance/dr-claw-plugin/tree/main/plugins/inno-research-pipeline/skills/inno-experiment-analysis
Command: npx skills add https://github.com/t2ance/dr-claw-plugin --skill inno-experiment-analysis-t2ance

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Experimental results are often scattered across files and notes, making it hard to derive reliable conclusions and write a crisp Results section.

Core Features & Use Cases

  • Statistical analysis pipeline: load results from CSV/JSON or TensorBoard logs, compute descriptive statistics, perform significance tests, and report effect sizes.
  • Visualization & reporting: generate vector-format figures and a ready-to-use Results draft that can be inserted into papers.
  • Workflow integration: coordinates data loading, validation, analysis, visualization, and writing to streamline paper generation.

Quick Start

Run the analysis workflow on your experiment results to generate an analysis-report.md, a results-draft.md, and a visualization-specs.md.

Frequently Asked Questions about inno-experiment-analysis

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

FAQPage Schema
How do I analyze ML experiment results and write a publication-ready Results section?

Analyze ML experiment results by loading CSV/JSON logs or TensorBoard curves, running statistical tests, and generating a ready-to-use draft Results section with publication-quality figures.

What statistical tests can I use to evaluate cross-model ablation performance?

Evaluate cross-model ablation performance using t-tests, ANOVA, or non-parametric options, and compute descriptive statistics alongside effect size calculations to validate experimental significance.

Does this experimental analysis workflow support TensorBoard curve data?

Yes, the experimental analysis workflow supports TensorBoard curve data alongside CSV and JSON logs, applying data validation to parse training curves for downstream statistical analysis.

How do I generate vector-format figures from machine learning experiment logs?

Generate vector-format figures from machine learning experiment logs by running the analysis workflow, which produces visualization specs and publication-quality charts directly from your validated data.

What's the best way to ensure reproducibility when translating ML experiments into insights?

Ensure reproducibility by applying the workflow's data validation during loading, computing standardized descriptive statistics, and outputting structured analysis reports alongside your visualization specs.

Can I produce a coherent cross-dataset ablation report from scattered experimental results?

Yes, you can produce a coherent cross-dataset ablation report by loading scattered experimental results, applying statistical significance tests across models, and generating a unified draft analysis report.