datacmp 3.0.0

A powerful Python library for data cleaning and exploratory data analysis

Datacmp

Datacmp is a powerful, lightweight Python library designed to simplify and accelerate data cleaning and exploratory data analysis (EDA) workflows. Built for data scientists and analysts, it provides intelligent preprocessing, structured insights, and beautiful visualizations—all with just a few lines of code.

---

Features

Smart Data Cleaning

• Automatic column name standardization (lowercase, underscores, no special chars)
• Intelligent missing value handling with configurable strategies (mean, median, mode)
• Outlier detection and handling using IQR method (cap or remove)
• Duplicate row removal with detailed logging
• Configurable threshold-based column dropping

Comprehensive Profiling

• Dataset overview with row/column counts and memory usage
• Detailed column analysis including dtypes, missing values, and unique counts
• Extended statistics (mean, median, std, skewness, kurtosis)
• Correlation analysis (Pearson, Spearman, Kendall)
• Column type detection (numeric, categorical, datetime)

Beautiful Visualizations

• Missing value heatmaps
• Correlation matrices
• Distribution plots for numeric features
• Export-ready plots in high resolution

Flexible Reporting

• HTML reports with interactive styling and embedded visualizations
• Text reports for quick inspection
• CSV export of cleaned datasets
• Complete audit trail of all cleaning operations

YAML-Based Configuration

• Fully decoupled configuration for reproducibility
• Pipeline versioning and easy sharing
• Template generation with datacmp init

Command-Line Interface

• Run pipelines directly from terminal
• Generate config files with defaults
• Progress tracking and verbose logging

---

Comparison: v2.0 → v3.0

Feature	v2.0	v3.0
API Style	Functional only	OOP + Functional
Method Chaining	❌	✅
Type Hints	Partial	Complete
Logging	print()	logging module
Visualizations	❌	✅ (3 types)
HTML Reports	❌	✅ (styled)
Correlations	❌	✅
CLI Subcommands	❌	✅
Test Coverage	None	Comprehensive
Packaging	setup.py	pyproject.toml

---

Installation

From PyPI (Recommended)

pip install datacmp

From Source

git clone https://github.com/MoustafaMohamed01/datacmp.git
cd datacmp
pip install -e .

With Optional Dependencies

# For full features
pip install datacmp[full]

# For development
pip install datacmp[dev]

---

Quick Start

Python API

from datacmp import DataCmp

# Load and process your data
cmp = DataCmp("data.csv")

# Clean, profile, and export in one chain
cmp.clean().profile().export("report.html")

# Or use individual methods
cmp = DataCmp("data.csv")
cmp.clean(outliers=True, duplicates=True)
cmp.profile(detailed=True)
cmp.visualize(output_dir="./plots")
cmp.export("cleaned_data.csv")

Command-Line Interface

# Run complete pipeline
datacmp run data.csv --config config.yaml --export cleaned.csv --report report.html

# Create default config file
datacmp init my_config.yaml

# Show version
datacmp version

---

Usage Examples

Example 1: Basic Cleaning

from datacmp import DataCmp

# Initialize with auto-clean
cmp = DataCmp("messy_data.csv", auto_clean=True)

# Export cleaned data
cmp.export("clean_data.csv")

# View cleaning log
print(cmp.get_cleaning_log())

Example 2: Custom Configuration

from datacmp import DataCmp

# Define custom config
config = {
    "cleaning": {
        "threshold_drop": 0.3,  # Drop columns with >30% missing
        "fill_strategy": {
            "numeric": "mean",
            "categorical": "mode"
        },
        "outlier_handling": {
            "enabled": True,
            "action": "remove"  # Remove outliers instead of capping
        }
    }
}

# Use custom config
cmp = DataCmp("data.csv", config=config)
cmp.clean().profile()

Example 3: Method Chaining

from datacmp import DataCmp

result = (
    DataCmp("data.csv")
    .clean(columns=True, missing=True, outliers=True)
    .profile(detailed=True)
    .visualize(output_dir="./plots")
    .export("report.html")
)

Example 4: Programmatic Pipeline

from datacmp import run_pipeline

# Run entire pipeline with one function
df_cleaned = run_pipeline(
    data="data.csv",
    config_path="config.yaml",
    export_csv_path="cleaned.csv",
    export_report_path="report.html"
)

---

Configuration

Example config.yaml

library_name: datacmp
version: 3.0.0
author: Moustafa Mohamed

cleaning:
  threshold_drop: 0.45
  fill_strategy:
    numeric: median
    categorical: mode
  outlier_handling:
    enabled: true
    method: iqr
    iqr_multiplier: 1.5
    action: cap

drop_duplicates: true

profiling:
  include_more_stats: true
  compute_correlations: true

Configuration Options

Option	Description	Default
threshold_drop	Drop columns with missing ratio above this	0.45
fill_strategy.numeric	Strategy for numeric columns (mean, median, mode)	median
fill_strategy.categorical	Strategy for categorical columns (mode)	mode
outlier_handling.enabled	Enable outlier detection	true
outlier_handling.method	Detection method (iqr)	iqr
outlier_handling.action	Action to take (cap, remove)	cap
drop_duplicates	Remove duplicate rows	true

---

Visualizations

Datacmp automatically generates:

• Missing Value Heatmaps - Visualize patterns in missing data
• Correlation Heatmaps - Identify relationships between features
• Distribution Plots - Understand feature distributions

cmp = DataCmp("data.csv")
cmp.clean().profile()
cmp.visualize(output_dir="./plots")

---

Reports

HTML Reports

Beautiful, interactive HTML reports with:

• Dataset overview and statistics
• Cleaning operation log
• Embedded visualizations
• Responsive design

cmp.export("report.html")

Text Reports

Lightweight text reports for quick inspection:

cmp.export("report.txt")

---

API Reference

DataCmp Class

DataCmp(data, config=None, auto_clean=False)

Methods:

• clean(columns=True, missing=True, outliers=True, duplicates=True) - Clean the dataset
• profile(detailed=True) - Generate profiling information
• visualize(output_dir=None) - Create visualizations
• export(output, format=None, include_plots=True) - Export results
• reset() - Reset to original DataFrame
• get_summary() - Get dataset summary string
• get_cleaning_log() - Get list of cleaning operations

---

Contributing

Contributions are welcome! Please read CONTRIBUTING.md for guidelines.

Development Setup

git clone https://github.com/MoustafaMohamed01/datacmp.git
cd datacmp
pip install -e ".[dev]"

---

Changelog

See CHANGELOG.md for version history.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Author

Moustafa Mohamed

• Email: [e-mail]
• Linkedin: Moustafa Mohamed
• GitHub: MoustafaMohamed01
• Kaggle: moustafamohamed01

---

Show Your Support

Give a ⭐️ if this project helped you!

---

Resources

• Documentation
• Issue Tracker
• PyPI Package