Skip to content

README.md

Latest commit

 

History

History
232 lines (163 loc) · 6.17 KB

README.md

File metadata and controls

232 lines (163 loc) · 6.17 KB
Raw

🚀 Ak-dskit Demos

Welcome to the comprehensive demo collection for Ak-dskit! This folder contains 12 detailed demonstration scripts showcasing all major functionalities of the library.

📚 Demo Overview

Each demo file is self-contained and demonstrates specific features of dskit with clear examples and explanations.

Core Functionality Demos

  1. 01_data_io_demo.py - Data Input/Output Operations

    • Load data from CSV, Excel, JSON, Parquet
    • Batch loading from folders
    • Save data in multiple formats
    • Smart data type detection

  2. 02_data_cleaning_demo.py - Data Cleaning

    • Automatic data type fixing
    • Column name standardization
    • Special character replacement
    • Missing value analysis and imputation
    • Outlier detection and removal
    • Text/NLP cleaning

  3. 03_eda_demo.py - Exploratory Data Analysis

    • Basic statistics
    • Quick EDA overview
    • Comprehensive EDA with visualizations
    • Data health checks
    • Feature analysis reports

  4. 04_visualization_demo.py - Data Visualization

    • Missing value patterns
    • Distribution histograms
    • Boxplots for outliers
    • Correlation heatmaps
    • Pairplots

  5. 05_preprocessing_demo.py - Data Preprocessing

    • Automatic categorical encoding
    • Feature scaling (Standard, MinMax, Robust)
    • Train-test splitting
    • Complete preprocessing pipeline

  6. 06_modeling_demo.py - Machine Learning Modeling

    • Quick model training
    • Model comparison
    • Hyperparameter optimization
    • Model evaluation
    • Error analysis

Advanced Feature Demos

  1. 07_feature_engineering_demo.py - Feature Engineering

    • Polynomial features
    • Date feature extraction
    • Binning and discretization
    • Univariate feature selection
    • Recursive Feature Elimination (RFE)
    • Principal Component Analysis (PCA)
    • Aggregation features
    • Target encoding

  2. 08_nlp_demo.py - NLP Utilities

    • Text statistics
    • Advanced text cleaning
    • Text feature extraction
    • Sentiment analysis
    • Complete NLP pipeline

  3. 09_advanced_visualization_demo.py - Advanced Visualization

    • Feature importance plots
    • Target distribution analysis
    • Feature vs target relationships
    • Advanced correlation analysis
    • Missing pattern visualization
    • Outlier visualization

  4. 10_automl_demo.py - AutoML & Optimization

    • Default parameter spaces
    • Random search optimization
    • Grid search optimization
    • Bayesian optimization
    • Method comparison

  5. 11_hyperplane_demo.py - Hyperplane Visualization

    • Hyperplane class usage
    • SVM hyperplane visualization
    • Logistic regression hyperplane
    • Hyperplane extraction
    • Algorithm comparison

  6. 12_complete_pipeline_demo.py - End-to-End Pipeline

    • Complete ML workflow
    • Data loading → Cleaning → EDA → Feature Engineering
    • Preprocessing → Modeling → Evaluation → Interpretation
    • Best practices demonstration

🚀 Quick Start

Run All Demos

# Navigate to demos folder
cd demos

# Run individual demos
python 01_data_io_demo.py
python 02_data_cleaning_demo.py
# ... and so on

# Or run all demos at once
python run_all_demos.py

Run Specific Demo

# Example: Run data cleaning demo
python 02_data_cleaning_demo.py

Interactive Usage

# Import and use demo functions interactively
from demos.demo_01_data_io import demo_basic_loading
demo_basic_loading()

📋 Requirements

All demos use the standard dskit installation:

pip install Ak-dskit

Some advanced features may require additional packages:

  • textblob for sentiment analysis
  • hyperopt or optuna for advanced optimization

🎯 Learning Path

Beginners: Start with demos 1-6 (Core Functionality)

  1. Data I/O → Data Cleaning → EDA
  2. Visualization → Preprocessing → Modeling

Intermediate: Continue with demos 7-9 (Advanced Features) 3. Feature Engineering → NLP → Advanced Visualization

Advanced: Explore demos 10-12 (AutoML & Pipelines) 4. AutoML → Hyperplane → Complete Pipeline

📊 Demo Features

Each demo includes:

  • ✅ Clear section headers and descriptions
  • ✅ Step-by-step explanations
  • ✅ Sample data generation
  • ✅ Multiple use case examples
  • ✅ Output interpretation
  • ✅ Best practices
  • ✅ Error handling examples

💡 Tips

  1. Read the code: Each demo is well-commented and self-explanatory
  2. Modify parameters: Experiment with different settings
  3. Use sample data: All demos create their own sample datasets
  4. Check outputs: Each demo prints detailed progress and results
  5. Visualizations: Some demos create plots (saved as temporary files)

🔧 Customization

All demos can be easily customized:

# Example: Modify demo parameters
from demos.demo_06_modeling import demo_compare_models

# Use your own data
import pandas as pd
df = pd.read_csv('your_data.csv')

# Customize and run
demo_compare_models(df, target_col='your_target')

📖 Documentation

For detailed API documentation, see:

🤝 Contributing

To add a new demo:

  1. Follow the existing demo structure
  2. Include comprehensive examples
  3. Add clear documentation
  4. Update this README

📞 Support

📝 License

These demos are part of the Ak-dskit package and follow the same license.

Happy Learning! 🎉

Start with 01_data_io_demo.py and work your way through to 12_complete_pipeline_demo.py for a complete understanding of dskit capabilities!