Introduction

Texturizer is a Python package which aims to provide an easy and intuitive way of generating features from columns of text in a dataset.

The current implementation has been developed in Python 3 and tested on a variety of CSV files.

Motivation

Text data can add significant value to machine learning projects, but it is not always obvious how to make use of it. There are a vast number of ways to exploit text as features in a model and it is not always clear what is likely to work.

This package is intended to provide a quick, as well as easily extensible framework to add columns to a dataset using a wide variety of feature engineering approaches.

It can be as either a CLI utility to process a tabular dataset, or as python package that can be included within your ML projects. We include a SciKit Learn Compatible Transformer for using in machine learning pipelines.

Limitations

• The majority of the features are achieved via RegEx patterns. This makes the features fast to calculate and easily extensible. But it is a very manual process.

• Currently the package supports English only. But this could be changed by swapping out the word patterns and dictionaries, and introducing alternative SpacY language models.