skills.rest
raphaelmansuy
raphaelmansuyCOMMUNITY
Explore all skills by this author →
Website𝕏@raphaelmansuyGitHub

sota-data-cleaning-feature-selection-eda

Automate EDA, cleaning, and feature selection for tabular datasets with Polars.

Updated Dec 26, 2025
Search Tags:#EDA#data-cleaning#feature-selection#Polars#Sweetviz#Colab#Kaggle
One-click install
npx skills add https://github.com/raphaelmansuy/machine-learning-feature-selection --skill sota-data-cleaning-feature-selection-eda
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sota-data-cleaning-feature-selection-eda
Source: https://github.com/raphaelmansuy/machine-learning-feature-selection/tree/main/.github/skills/sota-data-cleaning-feature-selection-eda
Command: npx skills add https://github.com/raphaelmansuy/machine-learning-feature-selection --skill sota-data-cleaning-feature-selection-eda

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates data preparation workflows for medium tabular datasets by combining automated EDA, cleaning, and feature selection to accelerate modeling.

Core Features & Use Cases

  • Automated EDA with Sweetviz to surface distributions, correlations, and leakage indicators.
  • Automated cleaning with Pyjanitor to fix naming, deduplicate, normalize, and impute missing values.
  • Feature selection using a hybrid approach (filters + embedded methods like XGBoost or Lasso) to reduce dimensionality before training.
  • Colab-friendly execution on datasets in the 100MB–5GB range with Polars for fast I/O.

Quick Start

  1. Install required packages in Colab: polars, sweetviz, pyjanitor, xgboost, scikit-learn.
  2. Load a medium dataset (approx 100MB–5GB) and run an automated EDA pass.
  3. Apply automated cleaning transforms and run a hybrid feature selection to prepare data for model training.

Frequently Asked Questions about sota-data-cleaning-feature-selection-eda

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

FAQPage Schema
How do I automate data preparation and feature selection for tabular datasets in Colab?▼

Automate data preparation in Colab by running automated EDA with Sweetviz, cleaning with Pyjanitor, and hybrid feature selection using XGBoost to achieve cross-validated model readiness for tabular datasets.

Can I use Polars for fast I/O on medium-sized datasets during EDA and data cleaning?▼

Polars supports fast I/O for medium-sized datasets during EDA and data cleaning, handling files in the 100MB to 5GB range efficiently within Colab to accelerate modeling workflows.

What is the best way to detect data leakage and reduce dimensionality before model training?▼

Detect data leakage and reduce dimensionality by running automated EDA to surface correlations and applying hybrid feature selection combining statistical filters with embedded methods like XGBoost.

Does this data cleaning workflow handle noisy data and missing value imputation automatically?▼

Data cleaning handles noisy data and missing values automatically using Pyjanitor to fix naming, deduplicate, normalize, and impute missing values before running feature selection.

How do I prepare Kaggle-style datasets for modeling using automated EDA and statistical filters?▼

Prepare Kaggle-style datasets by integrating automated EDA with Sweetviz to surface distributions, applying cleaning transforms, and using statistical filters plus embedded methods for cross-validated model readiness.

When should I use a hybrid feature selection approach instead of standard statistical filters?▼

Use a hybrid feature selection approach combining statistical filters with embedded methods like XGBoost or Lasso when you need robust dimensionality reduction before training on complex tabular datasets.

Related Skills

Explore More Data & Analytics Skills→
danielbodnardanielbodnar
3

nushell-data-processing

Process large datasets with Polars DataFrames and LazyFrames in Nushell.

Community
Intermediate
pur3v4d3rpur3v4d3r
6

scikit-learn

Train and evaluate classical machine learning models with scikit-learn.

Community
Advanced
coolsocketcoolsocket
1

data-analysis

Analyze datasets with pandas, polars, seaborn, and matplotlib for cleaning, EDA, visualization, and modeling.

Community
Advanced
Mohamed-ElkahefMohamed-Elkahef

data-analyst

Clean missing values in CSV datasets and generate Plotly Dash dashboards.

Community
Intermediate