Cognia

Automated Exploratory Data Analysis

Cognia is a Python library that automatically performs Exploratory Data Analysis (EDA) and generates a structured, insight-rich HTML report.

Instead of writing repetitive and error-prone EDA scripts, Cognia thinks like a data analyst and delivers clear insights, visualizations, and warnings instantly.

✨ Why Cognia?

Before building:

• Machine Learning models
• Statistical analyses
• Dashboards or business insights

You must understand your data.

However, traditional EDA is often:

❌ Time-consuming

❌ Repetitive

❌ Hard to standardize

❌ Difficult to scale

👉 Cognia automates this entire process.

📁 Project Structure:

COGNIA/
│
├── cognia/                     # Core Cognia package
│   ├── __init__.py             # Package initializer
│   ├── alert.py                # Data quality alerts & warnings
│   ├── corr.py                 # Correlation analysis utilities
│   ├── interpret.py            # Distribution & insight interpretation
│   ├── missing.py              # Missing value analysis
│   ├── outliers.py             # Outlier detection logic
│   ├── profiling.py            # Dataset profiling helpers
│   ├── quick_eda.py             # Fast high-level EDA summary
│   ├── report.py               # HTML report generation engine
│   └── stats.py                # Statistical computations
│
├── demo/                       # Demo & example files
│   ├── cognia_eda_report.html  # Sample generated EDA report
│   ├── input_file.py           # Example usage script
│   └── labtoprice.csv          # Sample dataset
│
├── pyproject.toml              # Build & dependency configuration
├── README.md                   # Project documentation

🔍 What Cognia Analyzes:

Cognia generates a complete EDA report covering:

📊 Dataset Overview:

• Total rows & columns
• Data types
• Duplicate records
• Numeric vs categorical features

❓ Missing Value Analysis:

• Column-wise missing counts
• Missing percentages
• Data completeness indicators

📈 Statistical Summary:

• Mean, median, standard deviation
• Min / Max values
• Distribution characteristics

📉 Distribution & Shape Analysis:

• Histograms for numeric features
• Skewness detection
• Interpretable insights

🚨 Outlier Detection:

• Outlier counts per column
• Severity-based alerts
• Early modeling risk detection

🧩 Categorical Feature Analysis:

• Top categories
• Frequency bar charts
• Color-coded visualizations

🔗 Correlation Analysis (Smart & Scalable):

• Top correlated feature pairs (for large datasets)
• Optional full correlation heatmap
• Human-readable layout (no clutter)

⚠️ Alerts & Warnings:

• High missing values
• Duplicate data risks
• Extreme skewness & outliers
• Potential modeling issues

🧪 How to Use Cognia?

from cognia import eda_report

eda_report(df)

✔️ That’s it.

✔️ An HTML EDA report is generated instantly.

✔️ No configuration required.

📦 Installation:

Clone the repository and install locally:

pip install -e .

🛠 Built With:

🐍 Python 3.8+

📦 pandas

🔢 numpy

📊 matplotlib

📂 HTML

Lightweight • Fast • Beginner-friendly • Extensible

🏁 Philosophy:

If you can load a DataFrame, you should be able to understand it.

Cognia makes that possible.

If you find Cognia useful, don’t forget to ⭐ star the repository and share it with fellow data enthusiasts.