A Quality-of-Life Data Science Helper library for common ML/DS tasks
Project description
DSHelper 🚀
A Quality-of-Life Data Science Helper Library
DSHelper is a comprehensive toolkit designed to eliminate repetitive boilerplate code in data science and machine learning workflows. It wraps common operations into simple, intuitive functions that save time and improve code readability.
✨ Features
- 🔍 Missing Value Analysis: Comprehensive tools for detecting, visualizing, and handling missing data
- 📊 Correlation Analysis: Quick correlation matrices, heatmaps, and feature selection
- ⚙️ Preprocessing Pipeline: Train-test split with automatic scaling in one line
- 📈 Model Evaluation: Instant metrics, confusion matrices, and visualizations
- 🎯 Feature Engineering: Outlier detection, encoding, and feature selection utilities
- 🔄 Cross-Validation: Easy cross-validation with beautiful summaries
📦 Installation
pip install dshelper
For development installation:
pip install dshelper[dev]
🚀 Quick Start
import pandas as pd
from dshelper import missing, preprocessing, correlation, evaluation
# Load your data
df = pd.read_csv('your_data.csv')
# 1. Analyze missing values
missing_report = missing.analyze(df, show_plot=True)
print(missing_report)
# 2. Quick correlation heatmap
corr_matrix = correlation.heatmap(df, method='pearson')
# 3. Split and scale in one go
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(
X, y, test_size=0.2, scaler='standard'
)
# 4. Train your model
from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier()
model.fit(X_train, y_train)
# 5. Quick evaluation
y_pred = model.predict(X_test)
metrics = evaluation.quick_eval(y_test, y_pred)
📚 Documentation
Missing Values Module
missing.analyze(df, threshold=0.0, show_plot=True)
Analyze missing values and generate a comprehensive report with visualization.
from dshelper import missing
# Get detailed missing value report
report = missing.analyze(df, threshold=5.0, show_plot=True)
# Shows only columns with >5% missing values
missing.quick_summary(df)
Get a quick statistical summary of missing values.
summary = missing.quick_summary(df)
print(f"Total missing: {summary['total_missing']}")
print(f"Missing percentage: {summary['missing_percentage']:.2f}%")
missing.fill_missing(df, strategy='mean')
Fill missing values using various strategies.
# Fill with mean
df_filled = missing.fill_missing(df, strategy='mean')
# Fill with median
df_filled = missing.fill_missing(df, strategy='median')
# Fill with constant
df_filled = missing.fill_missing(df, strategy='constant', fill_value=0)
Correlation Module
correlation.heatmap(df, method='pearson')
Generate beautiful correlation heatmaps.
from dshelper import correlation
# Basic heatmap
corr_matrix = correlation.heatmap(df)
# Spearman correlation with custom styling
corr_matrix = correlation.heatmap(
df,
method='spearman',
figsize=(14, 12),
threshold=0.5 # Only show correlations > 0.5
)
correlation.top_correlations(df, target='price', n=10)
Find features most correlated with a target variable.
# Get top 10 features correlated with target
top_features = correlation.top_correlations(df, target='price', n=10)
print(top_features)
correlation.remove_highly_correlated(df, threshold=0.95)
Remove multicollinear features.
df_reduced, removed_cols = correlation.remove_highly_correlated(
df, threshold=0.95
)
print(f"Removed columns: {removed_cols}")
Preprocessing Module
preprocessing.split_and_scale(X, y, test_size=0.2, scaler='standard')
Split data and apply scaling in one step.
from dshelper import preprocessing
# Standard scaling (default)
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(
X, y, test_size=0.2, scaler='standard', random_state=42
)
# MinMax scaling
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(
X, y, scaler='minmax'
)
# Robust scaling (good for outliers)
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(
X, y, scaler='robust'
)
preprocessing.encode_categorical(df, method='onehot')
Encode categorical variables easily.
# One-hot encoding
df_encoded = preprocessing.encode_categorical(
df, columns=['category', 'type'], method='onehot'
)
# Label encoding
df_encoded = preprocessing.encode_categorical(
df, method='label'
)
preprocessing.handle_outliers(df, method='iqr', action='remove')
Detect and handle outliers.
# Remove outliers using IQR method
df_clean = preprocessing.handle_outliers(df, method='iqr', action='remove')
# Clip outliers instead of removing
df_clipped = preprocessing.handle_outliers(df, method='iqr', action='clip')
# Flag outliers (adds boolean columns)
df_flagged = preprocessing.handle_outliers(df, action='flag')
Evaluation Module
evaluation.quick_eval(y_true, y_pred, task_type='auto')
Comprehensive model evaluation in one line.
from dshelper import evaluation
# Auto-detect classification or regression
metrics = evaluation.quick_eval(y_test, y_pred)
# Prints detailed metrics and shows visualizations:
# - Classification: accuracy, precision, recall, F1, confusion matrix
# - Regression: R², RMSE, MAE, residual plots
evaluation.compare_models(results, metric='accuracy')
Compare multiple models side by side.
results = {
'Logistic Regression': {'accuracy': 0.85, 'f1': 0.83, 'precision': 0.84},
'Random Forest': {'accuracy': 0.88, 'f1': 0.86, 'precision': 0.87},
'XGBoost': {'accuracy': 0.90, 'f1': 0.89, 'precision': 0.90}
}
comparison_df = evaluation.compare_models(results, metric='accuracy')
evaluation.cross_val_summary(model, X, y, cv=5)
Easy cross-validation with visualization.
from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier()
cv_results = evaluation.cross_val_summary(
model, X, y, cv=5,
scoring=['accuracy', 'f1', 'precision']
)
evaluation.feature_importance_plot(model, feature_names)
Visualize feature importances for tree-based models.
# After training a tree-based model
importance_df = evaluation.feature_importance_plot(
model, feature_names=X_train.columns.tolist(), top_n=20
)
🎯 Real-World Example
Here's a complete workflow using DSHelper:
import pandas as pd
from dshelper import missing, correlation, preprocessing, evaluation
from sklearn.ensemble import RandomForestClassifier
# 1. Load and explore data
df = pd.read_csv('titanic.csv')
# 2. Check missing values
missing_report = missing.analyze(df, show_plot=True)
# 3. Handle missing values
df_clean = missing.fill_missing(df, strategy='median', columns=['Age'])
df_clean = missing.fill_missing(df_clean, strategy='mode', columns=['Embarked'])
# 4. Check correlations
correlation.heatmap(df_clean)
top_features = correlation.top_correlations(df_clean, target='Survived', n=5)
# 5. Encode categorical variables
df_encoded = preprocessing.encode_categorical(
df_clean,
columns=['Sex', 'Embarked'],
method='onehot'
)
# 6. Prepare features and target
X = df_encoded.drop('Survived', axis=1)
y = df_encoded['Survived']
# 7. Split and scale
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(
X, y, test_size=0.2, scaler='standard', stratify=True
)
# 8. Train model
model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)
# 9. Evaluate
y_pred = model.predict(X_test)
metrics = evaluation.quick_eval(y_test, y_pred)
# 10. Cross-validation
cv_results = evaluation.cross_val_summary(model, X_train, y_train, cv=5)
# 11. Feature importance
importance = evaluation.feature_importance_plot(model, X.columns.tolist())
🛠️ Requirements
- Python >= 3.8
- numpy >= 1.20.0
- pandas >= 1.3.0
- scikit-learn >= 1.0.0
- matplotlib >= 3.3.0
- seaborn >= 0.11.0
🤝 Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
- Fork the repository
- Create your feature branch (
git checkout -b feature/AmazingFeature) - Commit your changes (
git commit -m 'Add some AmazingFeature') - Push to the branch (
git push origin feature/AmazingFeature) - Open a Pull Request
📝 License
This project is licensed under the MIT License - see the LICENSE file for details.
📧 Contact
Ayush Lokre - ayushlokre5@gmail.com
Project Link: https://github.com/ayushlokre/dshelper
🙏 Acknowledgments
- Built with ❤️ for the data science community
- Inspired by the repetitive nature of data science workflows
- Thanks to all contributors and users
📊 Why DSHelper?
Before DSHelper:
# Multiple lines of repetitive code
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)
With DSHelper:
# One clean line
X_train, X_test, y_train, y_test = preprocessing.split_and_scale(X, y)
Happy Data Science! 🎉
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file dshelper_ayushlokre-0.1.0.tar.gz.
File metadata
- Download URL: dshelper_ayushlokre-0.1.0.tar.gz
- Upload date:
- Size: 21.6 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
afb0dd544afc87add802860645858844102dbc99b64c24859a9e7f4f9ff4f21e
|
|
| MD5 |
7c1ef70219f1cd77befdd80f578c2a78
|
|
| BLAKE2b-256 |
17a9574e706ed2dea46d802e58a1f02013b4e58fb58107ded881bcd48ad0a16d
|
File details
Details for the file dshelper_ayushlokre-0.1.0-py3-none-any.whl.
File metadata
- Download URL: dshelper_ayushlokre-0.1.0-py3-none-any.whl
- Upload date:
- Size: 19.5 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
dcc94d1df0ecec16ac2eedac2b0388fc1192a70a482f17bb4ad6b34c8e851882
|
|
| MD5 |
acf99bb8cc7af7b1996867a093720ba8
|
|
| BLAKE2b-256 |
0e8aeefd9a2d8c9b25510eeab9bfcc1748d3cb1c5506c783bd38c3ef0d0d873a
|