skills.rest
DuqueOM
DuqueOMCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

eda-analysis

Run a six-phase EDA pipeline on raw datasets to generate training, schema, and drift artifacts.

5|Updated Nov 18, 2025
Search Tags:#drift-detection#eda#data-quality#data-processing#feature-engineering#exploratory-data-analysis#schema-proposal
One-click install
npx skills add https://github.com/DuqueOM/ML-MLOps-Portfolio --skill eda-analysis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: eda-analysis
Source: https://github.com/DuqueOM/ML-MLOps-Portfolio/tree/main/.devin/skills/eda-analysis
Command: npx skills add https://github.com/DuqueOM/ML-MLOps-Portfolio --skill eda-analysis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill guides an agent through a six-phase exploratory data analysis pipeline that surfaces data-centric artifacts used in model training, schema design, and production drift monitoring.

Core Features & Use Cases

  • A six-phase EDA workflow that ingests, profiles, analyzes univariate distributions, surfaces correlations, checks for leakage, and proposes feature candidates.
  • Produces artifacts consumed by training (features.py), schema generation (schemas.py), and drift detection (baseline_distributions.parquet).
  • Onboard new datasets into ML pipelines and validate data quality before modeling.

Quick Start

Ingest a dataset placed in data/raw and run the six-phase EDA pipeline to generate all artifacts and reports.

Frequently Asked Questions about eda-analysis

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

FAQPage Schema
How do I perform exploratory data analysis on a new dataset for machine learning?▼

Exploratory data analysis profiles univariate distributions, surfaces correlations, and checks for data leakage. This Skill automates a six-phase EDA pipeline that ingests raw datasets to generate training artifacts, schema designs, and baseline distributions for drift detection.

How do I generate baseline distributions for production drift detection?▼

To generate baseline distributions for drift detection, run the EDA pipeline on your raw data. It produces a baseline_distributions.parquet artifact capturing the statistical profile needed to monitor production data drift.

What is the best way to propose feature candidates and audit data leakage automatically?▼

Auditing data leakage and proposing feature candidates is best handled by an end-to-end EDA pipeline. This workflow assesses correlations and validates data quality, outputting a features.py file with proposed candidates for model training.

Can I generate a schema proposal directly from raw dataset profiling?▼

Yes, you can generate a schema proposal from raw dataset profiling. The EDA workflow analyzes ingested data and emits a schemas.py artifact defining structure and constraints for downstream schema design.

Do I need DVC to track exploratory data analysis artifacts?▼

You need DVC to track EDA artifacts if you require enforced governance checks. The workflow operates on data/raw datasets and enforces DVC-tracked artifacts to manage outputs generated during the six-phase analysis.

Related Skills

Explore More Data & Analytics Skills→
xjtulycxjtulyc
33

data-version-control

Version datasets and orchestrate reproducible pipeline stages with DVC.

Community
Intermediate
ThejanaJayalathThejanaJayalath

mle-workflow

Standardize machine learning lifecycle from data contracts to production monitoring.

Community
Advanced
HermeticOrmusHermeticOrmus

data-versioning-patterns

Implement reproducible ML data management with DVC, Delta Lake, and OpenLineage.

Community
Advanced
TeoSlayerTeoSlayer
7

pilot-data-labeling-pipeline-setup

Deploy a four-agent data labeling pipeline with ingestion, auto-labeling, review, and export.

Community
Advanced