User Guide

Comprehensive guides for using edaflow effectively in your data analysis and machine learning workflows.

• Learning Path for Data Science with edaflow
  • Step 1: Getting Started
  • Step 2: Data Quality & Cleaning
  • Step 3: Visualization & Analysis
  • Step 4: Machine Learning Workflows
  • Step 5: Advanced Features & Best Practices
  • Step 6: API Reference & Further Exploration
• Data Quality & Cleaning
  • Overview
• Advanced Time Series Topics
• 1. Forecasting
• 2. Autocorrelation & Lag Analysis
• 3. Feature Engineering for Time Series
• 4. Integrating Time Series Models
  • Visualization Guide
• Getting Started
• Correlation Analysis
• Advanced & Interactive Plots
• More Visualization Examples
• Best Practices
  • External Library Requirements for Advanced Features
• Feature Dependency Table:
• How edaflow Splits Your Dataset: Training, Validation, and Test
• Further Resources & FAQ
• Choosing the Right Performance Visualization
• Overview
• Data Validation: A Critical First Step
• Best practice: Aim for a high data quality score to ensure robust, reliable model results.
  • About cv_folds
  • Other Key Parameters for Model Comparison
  • Choosing Metrics by Problem Type
  • Practical Examples: Metric Selection
• Complete ML Workflow Example
• Individual Function Examples
  • Configuration Functions
  • This workflow helps you make informed, data-driven choices about which model to use in production or further tuning.
  • Hyperparameter Tuning Functions
• Why is it important?
• How does it work?
  • Performance Visualization Functions
  • Model Artifacts Functions
• Best Practices
• This approach ensures you have a solid reference point and helps you build more robust, trustworthy machine learning solutions.
• Widely Used Model Types in Machine Learning
• Refer to scikit-learn and the respective library documentation for more details and advanced options.
• What’s Next After Training the Model?
• Machine Learning Workflow with edaflow
  • Step 1: Setup ML Experiment
  • Step 2: Compare Multiple Models
  • Step 3: Optimize Hyperparameters
  • Step 4: Rank and Select Best Model
  • Step 5: Save Model Artifacts
  • Step 6: Visualize Learning Curves
• Advanced Features in edaflow
  • New Advanced Features
• Best Practices for edaflow
  • Best Practices for New Features

Overview

The edaflow User Guide is organized into five main sections:

Data Quality & Cleaning

Learn how to assess data quality, handle missing values, convert data types, and prepare your data for analysis.

• Missing data analysis and visualization

• Categorical data insights and type conversion

• Data imputation strategies

• Outlier detection and handling

Visualization & Analysis

Explore edaflow’s comprehensive visualization capabilities for understanding your data.

• Distribution analysis with boxplots and histograms

• Interactive visualizations with Plotly

• Correlation and relationship analysis

• Advanced scatter matrix analysis

Machine Learning Workflows

Master the complete ML pipeline with edaflow’s comprehensive machine learning functions.

• ML experiment setup and data validation

• Multi-model comparison and ranking systems

• Hyperparameter optimization strategies

• Performance visualization and model artifacts

• Complete workflow examples and best practices

Advanced Features

Discover advanced features and customization options for power users.

• Custom thresholds and parameters

• Integration with other libraries

• Performance optimization tips

• Extension and customization

Best Practices

Learn recommended workflows and best practices for effective EDA and ML.

• Recommended EDA workflow

• Memory and performance considerations

• Jupyter notebook integration

• Troubleshooting common issues

New Features & Advanced Visualization

edaflow now supports: - Faceted visualizations with display_facet_grid - Feature scaling with scale_features - Grouping rare categories with group_rare_categories - Exporting figures with export_figure

See the Visualization Guide and Advanced Features for details and examples.

External Library Requirements

Some features require additional libraries (seaborn, scikit-learn, statsmodels). See the Visualization Guide for installation instructions and troubleshooting tips.

Getting Started

If you’re new to edaflow, we recommend starting with the Quick Start Guide guide, then exploring each section of this user guide based on your specific needs.

For complete function documentation, see the API Reference.