deep-study 1.0.3

Deep Study - Dataset Overview and Feature Analysis for Data Science

Deep Study

Automated Dataset Overview and Feature Analysis for Data Science

Deep Study is a lightweight Python package for automated exploratory data analysis (EDA), featuring detailed feature profiling and analysis with target variables. Generate professional HTML reports with just a few lines of code.

Features

• Feature Profiling: Detailed statistics for each feature (distinct values, missing values, distribution, memory usage)
• Target Analysis: Automatic analysis of feature relationships with target variable
• Feature Importance: ML-based feature importance using Random Forest
• Beautiful Reports: Generate professional HTML reports with visualizations
• Jupyter Integration: Reports render beautifully in Jupyter notebooks
• Easy to Use: Simple API - just 3 lines of code to generate a complete analysis

Installation

pip install deep-study

Or install from source:

git clone https://github.com/arshadziban/deep_study_lib.git
cd deep_study_lib
pip install -e .

Quick Start

from deep_study import Analyzer

# Create analyzer with your data
analyzer = Analyzer(df, target="target_column")

# Run analysis
report = analyzer.run()

# Display in Jupyter notebook (automatically renders HTML)
report

# Or save HTML report to file
report.save_html("deep_study_report.html")

# Print summary to console
report.summary()

Report Contents

The generated HTML report includes:

1. Dataset Summary

• Total rows and columns
• Target variable information
• Overall dataset statistics

2. Feature Overview with Target Variable

• All features with their types and statistics
• Relationship indicators with target variable
• Missing value analysis

3. Individual Feature Profiles

For each feature in your dataset:

• Type: Numeric, Categorical, or DateTime
• Distinct Values: Count and percentage
• Missing Values: Count and percentage
• Memory Usage: Efficient memory tracking
• Distribution Visualization: Histograms for numeric, bar charts for categorical
• Statistics: Mean, min, max (with K/M/B formatting for large numbers)
• Top Values: Most frequent values for categorical features

API Reference

Analyzer

from deep_study import Analyzer

analyzer = Analyzer(data, target)

Parameters:

• data: pandas DataFrame or path to CSV/Excel file
• target: Name of the target column

Methods:

• run(): Execute analysis and return Report object

Report

Methods:

• save_html(filename): Save report to HTML file
• summary(): Print analysis summary to console
• get_top_features(n): Get top N important features
• get_feature_profile(name): Get profile for specific feature

Example Output

from deep_study import Analyzer
import pandas as pd

# Load your data
df = pd.read_csv("your_data.csv")

# Analyze
analyzer = Analyzer(df, target="outcome")
report = analyzer.run()

# View in notebook or save
report.save_html("analysis_report.html")

Requirements

• Python >= 3.8
• pandas >= 1.3.0
• numpy >= 1.20.0
• scikit-learn >= 1.0.0
• matplotlib >= 3.4.0
• jinja2 >= 3.0.0

Author

Shah Md. Arshad Rahman Ziban

Acknowledgments

• Built with pandas, scikit-learn, and matplotlib
• Inspired by the need for quick, professional data analysis reports