viniciusmsousa / ds-toolbox Goto Github PK

A set of functions to help the analytical work of a data scientist. Full documentation can be found in Package Homepage. The main motivation of the package is to facilitate the taks by using a common input and output structure, SparkDF and PandasDF.

The package can be installed either using PyPi:

pip install ds-toolbox

Or directly form github:

pip install git+https://github.com/viniciusmsousa/ds-toolbox.git#main

• statistics:
  • contigency_chi2_test: Wrapper for Scipy function;
  • mannwhitney_pairwise: Wrapper for scikit-posthocs function;
  • ks_test: Compute the KS-Test, for Pandas and Spark DF;
  • ab_test_pairwise: A Simple ab test based on mean, std and var, PandasDF and Spark DF.
• ml:
  • evaluator:
    • binary_classifier_metrics: Computes classification metrics (confusion_matrix, accuracy, f1, precision, recall, aucroc, aucpr) based on a dataframe (SparkDF or PandasDF) with ground truth and prediction.
• econometrics:
  • causal_regression:
    • CausalRegression: A class built on top of what is presented in the chapters 19-21 from the book Causal Inference for The Brave and True.