New VerticaPy Features Added – Just in Time for the Holidays

Posted December 18, 2020 by Badr Ouali, Head of Data Science

Vertica wishes you a Happy Holidays

The team behind VerticaPy wishes you happy holidays and a happy new year from Boston, and from Grenoble in the French Alps!

Make 2021 your year to operationalize Machine Learning at Scale with Vertica and Python.  

 VerticaPy is a Python library that exposes scikit-like functionality for conducting data science projects in your Vertica database. With in-database machine learning, data preparation and exploration, VerticaPy lets you manipulate and analyze your data while leveraging the strength of Vertica – the most powerful unified analytics warehouse on the market. VerticaPy brings the logic to the data: everything happens in your database.

 Here some helpful new features in VerticaPy v.04:

  • Calculation of the Weight of Evidence (WOE) and Information Value (IV) – understand the influence of your predictors on your target
  • Support for SHAP Explainers – understand what influences your Vertica model
  • More Vertica model conversion – convert your Vertica model to an sklearn model
  • GeoPandas – draw beautiful maps
  • “Magic” methods – use even more Pandas-like functions for filtering data and computing features
  • Much more: