conversation-analytics-toolkit

Watson Assistant Dialog Flow Analysis

Note: help us stay in touch and improve this notebook by clicking on the :star: star icon (top right).

This repository hosts the the Watson Assistant Dialog Flow Analysis Notebook and the underlying conversation analytics toolkit library.

Table of Contents

Introduction
Getting Started
Guides
Frequently Asked Questions
License
Contributing

Introduction

The Watson Assistant Dialog Flow Analysis Notebook can help you assess and analyze user journeys and issues related to the dialog flow of ineffective (low quality) conversations based on production logs. The notebook can help you with questions such as:

• What are the common conversation steps and flows within the assistant
• Which flows have low task completion rates and high abandonment (ineffective conversations)
• Where along the dialog steps users lose engagement with your assistant
• What are common terms and steps that may lead to abandonment

This notebook extends the Measure and Analyze notebooks by providing additional capabilities to assess and analyze effectiveness - focused more on issues related to the dialog flow. For more details, check out IBM Watson Assistant Continuous Improvement Best Practices.

Getting Started

The notebook requires a Jupyter Notebook environment and Python 3.6+. You can either install Jupyter Notebook to run locally or you can use Watson Studio on the cloud.

Using Jupyter Notebook

1. Install Python 3.6+
2. Install Jupyter notebook. Checkout the Jupyter/IPython Notebook Quick Start Guide for more details
3. Download the notebooks/Dialog Flow Analysis Notebook.ipynb file.
4. Start jupyter server jupyter notebook
5. Run the Dialog Flow Analysis Notebook.ipynb

Using Watson Studio

1. In Watson Studio, select Add to Project-->Notebook. Choose From URL and paste this url. Alternately you can select From file and upload the notebooks/Dialog Flow Analysis Notebook.ipynb file.

Alternately, you can import and modify the sample notebook on Watson Studio Gallery.

Guides

• Learn more about the Dialog Flow Analysis in this blog
• See a live example output of the notebook on Watson Studio Gallery

Frequently Asked Questions

See FAQ.md for frequently asked questions

License

This library is licensed under the Apache 2.0 license.

Contributing

See CONTRIBUTING.md and DEVELOPER.MD for more details on how to contribute

Contributor List

Avi Yaeli	Sergey Zeltyn	Zhe Zhang	Eric Wayne	David Boaz