eda-datapilot 0.2.0

DataPilot - A powerful, automated Exploratory Data Analysis (EDA) library in Python

🚀 DataPilot: Automated Exploratory Data Analysis Library

datapilot is a modern, modular Python library for automated Exploratory Data Analysis (EDA). It generates comprehensive dataset overviews, missingness analysis, statistical summaries, outlier/duplicate detection, correlation matrices, automated data quality auditing, actionable suggestions, ML task detection, feature importance analysis, interactive charts, and standalone HTML reports.

---

🚀 Key Features

• Phase 1: Basic Summaries & Statistics
  • Dataset overview (rows, columns, memory, data types breakdown)
  • Missing value analysis (counts, percentages, visual progress bars)
  • Comprehensive numerical statistics (mean, median, mode, std, skewness, kurtosis, quantiles, zeros)
  • Categorical statistics (unique values, most frequent value, top 10 categories breakdown)
  • Duplicate detection (rows, %, constant columns, duplicate column pairs)
  • Correlation matrices (Pearson, Spearman, Kendall)
  • Outlier detection (IQR & Z-score bounds, counts, indices)
• Phase 2: Automatic Data Quality Audit
  • Instant alerts for high correlation, zero variance, high cardinality, missing values, class imbalance, and severe skewness.
• Phase 3: Actionable Suggestions
  • Automated recommendations for imputation strategies, log/power transformations, feature drops, and outlier clipping.
• Phase 4: Machine Learning Insights
  • Automatic task detection (Regression, Binary/Multiclass Classification, Time Series, Clustering)
  • Feature Importance calculation via Random Forests & Mutual Information
  • PCA variance ratio analysis
• Phase 5: Interactive HTML Report Generation
  • Standalone, self-contained HTML reports powered by Jinja2 & Plotly.

---

📦 Installation

pip install -e .

---

💻 Quick Start

Simple Functional API

import pandas as pd
import datapilot

# Load dataset
df = pd.read_csv("dataset.csv")

# Run full automated analysis & show report in browser
report = datapilot.analyze(df, target="target_column")
report.show("datapilot_report.html")

Object-Oriented API

from datapilot import EDA

eda = EDA(df, target="target_column")

# Step-by-step analytical calls
summary = eda.summary()
missing = eda.missing()
outliers = eda.outliers()
correlation = eda.correlation()
problems = eda.problems()
recs = eda.recommendations()

# Generate & save report
report = eda.report()
report.to_html("DataPilot_Report.html")

---

🛠️ Project Structure

datapilot/
│
├── datapilot/
│   ├── __init__.py
│   ├── analyzer.py
│   ├── summary.py
│   ├── missing.py
│   ├── numeric.py
│   ├── categorical.py
│   ├── correlation.py
│   ├── outliers.py
│   ├── duplicates.py
│   ├── problem_detector.py
│   ├── recommender.py
│   ├── ml_insights.py
│   ├── visualization.py
│   ├── report.py
│   ├── utils.py
│   ├── config.py
│   └── templates/
│       └── report.html.j2
│
├── tests/
│   ├── test_analyzer.py
│   ├── test_summary.py
│   ├── test_outliers.py
│   ├── test_report.py
│   └── test_recommendations.py
│
├── examples/
│   └── demo.py
├── pyproject.toml
├── setup.py
├── requirements.txt
└── README.md