Example for using Kedro for an NLP pipeline

In this example, Kedro was used to create a workflow to achieve the following:

• Pre-process the Sentiment140 dataset containing 1.6 million tweets
• Train a Logistic Regression model for sentiment prediction based on the text of the tweets
• Evaluate the trained model and generate a classification report

This project was created as part of the PyCon India 2020 Devsprint.

Workflow Visualization using kedro-viz

Acknowledgement

Special thanks to Lais Carvalho for her guidance through the devsprint.