autooutlier

Automatic Outlier Detection and Handling for Python

---

Overview

autooutlier is a Python package that automatically detects, analyzes, and handles outliers in numerical data. It intelligently selects the best detection and handling methods based on data distribution — requiring zero configuration from the user.

Simply pass your DataFrame and column name, and autooutlier handles the rest.

---

Features

• Automatic Detection — Selects the optimal outlier detection method (Z-Score, Modified Z-Score, IQR, Percentile) based on data skewness.
• Automatic Handling — Chooses the best outlier replacement strategy (winsorization, mean/median/mode replacement, interpolation, etc.) based on data characteristics.
• Statistical Analysis — Provides mean, median, mode, standard deviation, variance, skewness, kurtosis, and distribution classification.
• Pre-Cleaning Summary — Generates a comprehensive report before cleaning, including detection method, handling strategy, outlier count, and percentage.
• Post-Cleaning Report — Returns both the cleaned dataset and an after-cleaning summary report.
• Flexible Manual Control — Override automatic selections with manual detection and handling methods when needed.
• Visualization — Built-in box plot support via Seaborn.

---

Installation

pip install autooutlier

Or install from source:

git clone https://github.com/suruthika-cd/autooutlier.git
cd autooutlier
pip install -e .

Dependencies

• Python >= 3.8
• NumPy >= 1.21.0
• Pandas >= 1.3.0
• SciPy >= 1.7.0
• Seaborn >= 0.11.0
• Matplotlib >= 3.4.0

---

Quick Start

import pandas as pd
from autooutlier import handle_outliers, before_cleaning_summary, detect_outliers

# Load your data
df = pd.DataFrame({"values": [10, 12, 14, 11, 13, 100, 15, 12, 14, 11]})

# Get a pre-cleaning summary report
summary = before_cleaning_summary(df, "values")
print(summary)

# Automatically detect and handle outliers
cleaned_data, report = handle_outliers(df, "values")
print(report)
print(cleaned_data)

---

Usage Examples

Automatic Outlier Detection

from autooutlier import detect_outliers

outlier_mask = detect_outliers(df, "column_name")
print(f"Outliers found: {outlier_mask.sum()}")

Automatic Outlier Handling

from autooutlier import handle_outliers

# Fully automatic — detection and handling methods are chosen for you
cleaned_df, report = handle_outliers(df, "column_name")

Manual Detection Method

# Use a specific detection method
cleaned_df, report = handle_outliers(df, "column_name", detection_method="z_score")

Available detection methods: 'auto', 'Iqr_method', 'z_score', 'modified_z_score', 'percentile'

Manual Handling Method

# Use a specific replacement strategy
cleaned_df, report = handle_outliers(df, "column_name", replacement="median")

Available replacement methods: 'auto', 'interpolate', 'winsorization', 'median', 'mean', 'mode', 'custom', 'remove', 'bfill', 'ffill'

Custom Value Replacement

cleaned_df, report = handle_outliers(df, "column_name", replacement="custom", value=0)

Pre-Cleaning Summary

from autooutlier import before_cleaning_summary

summary = before_cleaning_summary(df, "column_name")
print(summary)

Output includes: suggested detection method, handling method, skewness, distribution type, outlier count, and outlier percentage.

---

API Overview

Public API

Function	Description
handle_outliers(data, column, detection_method='auto', replacement='auto', value=None)	Detect and handle outliers. Returns (cleaned_data, report).
detect_outliers(data, column)	Detect outliers automatically. Returns a boolean mask.
detect_outlier_method(data, column)	Returns the suggested detection method name.
before_cleaning_summary(data, column)	Returns a DataFrame summary report before cleaning.

Module Reference

Module	Contents
autooutlier.statistics	mean, median, mode, std, var, data_range, q1, q3, iqr, skew, skew_measurment, is_normal, kurtosis, kurtosis_measurement
autooutlier.detection	Iqr_method, z_score_method, modified_z_score, percentile_method, detect_outlier_method, detect_outliers
autooutlier.handling	winsorization, interpolate, replace_with_mean, replace_with_median, replace_with_mode, replace_with_custom_value, replace_with_forward_fill, replace_with_backward_fill, remove_outliers, detect_handler, handle_outliers
autooutlier.summary	before_cleaning_summary
autooutlier.visualization	box_plot
autooutlier.utils	is_numeric, is_time_series, is_continous, outlier_count, outlier_percentage

---

License

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

---

Changelog

See CHANGELOG.md for all notable changes.