Introducing Skippa

Summary

Every data scientist is probably familiar with pandas and scikit-learn. The usual workflow starts with cleaning data in pandas, followed by further pre-processing using pandas or scikit-learn transformers such as StandardScaler and OneHotEncoder. Then you move on to a machine learning algorithm (scikit-learn).

There are a few problems with this workflow:
1. The development phase of your workflow is quite complex and requires a lot of code ?
2. It is difficult to reproduce the workflow for predictions during deployment ?
3. Existing solutions to reduce these problems are not good enough (yet) ?

Skippa is a package designed to:

• ✨ drastically simplify development
• ? package all data cleaning and pre-processing together with the algorithm in a single pipeline file
• ? reuse the pandas and scikit-learn interfaces you already know"

Skippa helps you define data cleaning and pre-processing transformations easily. Here is roughly how it works:

from skippa import Skippa, columns from sklearn.linear_model import LogisticRegression X, y = get_training_data(...) pipeline = ( Skippa() .impute(columns(dtype_include='object'), strategy='most_frequent') .impute(columns(dtype_include='number'), strategy='median') .scale(columns(dtype_include='number'), type='standard') .onehot(columns(['category1', 'category2'])) .model(LogisticRegression()) ) pipeline.fit(X, y) predictions = pipeline.predict_proba(X)

☝️Skippa does not claim to solve every problem, cover every feature you might ever need or offer a highly scalable solution, but it should provide a substantial simplification for > 80% of typical pandas/sklearn-based machine learning projects. 

Links

• ? Github
• ? pypi
• ? readthedocs

Read the rest of the blog post here > Introduction Skippa [ENG]