A step-by-step customer propensity modelling experiment with Azure Machine Learning Studio

In this video, endjineer Ed Freeman performs a "worked example" customer propensity experiment using Azure Machine Learning Studio.

"Our hypothetical situation is based on an e-commerce business that is interested in changing its business processes to prioritise customers that are more likely to purchase their new service, based on the information they provide at various stages throughout the purchasing workflow."

The worked example covers:

- What is Customer Propensity? 0:20
- What are our Classification Problems? 0:35
- The Purpose of Walkthrough 1:02
- General Method 1:41
- What is our example? 2:45
- What is our Hypothesis? 3:05
- Interpretation of results​ 3:32
- Desired Metric - what is AUC? 4:02
- Worked Example in Azure Machine Learning Studio 4:26
- What did we conclude? 36:35

You can sign up for your free Azure Machine Learning Studio account (no credit card required) via this link. https://studio.azureml.net/ - it's very easy to get started.

Other endjin Machine Learning Blog Posts:

"Machine Learning – mad science or a pragmatic process?": https://blogs.endjin.com/2016/02/machine-learning-mad-science-or-a-pragmatic-process/

"Machine Learning – the process is the science": https://blogs.endjin.com/2016/03/machine-learning-the-process-is-the-science/

"Demystifying Machine Learning: using neural networks": https://blogs.endjin.com/2019/07/demystifying-machine-learning-using-neural-networks/

"Azure Machine Learning–experimenting with training data proportions using the SMOTE module": https://blogs.endjin.com/2015/09/azure-machine-learning-experimenting-with-training-data-proportions-using-the-smote-module/

"Using Postman to load test an Azure Machine Learning web service": https://blogs.endjin.com/2016/04/using-postman-to-load-test-an-azure-machine-learning-web-service/

"Using Python inside SQL Server": https://blogs.endjin.com/2018/01/using-python-inside-sql-server/

"ML.NET, Azure Functions and the 4th Industrial Revolution": https://blogs.endjin.com/2019/02/ml-net-azure-functions-and-the-4th-industrial-revolution/

You might also be interested in the free Jupyter Notebooks experience on Azure: https://notebooks.azure.com

We'd also strongly recommend taking a look at the underrated Azure ML DataPrep SDK which is available in both Python and .NET SDKs

https://docs.microsoft.com/en-us/dotnet/api/overview/data-prep/overview-data-prep?view=dataprep-dotnetcore

https://github.com/Azure-Samples/DataPrep.Net

https://docs.microsoft.com/en-gb/python/api/overview/azureml-sdk/?view=azure-ml-py#dataprep

Ed also curated the Power BI Weekly newsletter, which you can read, or subscribe to via https://powerbiweekly.info

You can follow Ed Freeman on twitter at https://twitter.com/edfreeman_

If you have any questions, please leave a comment and we'll do our best to answer them.