Data Science in the Cloud with Microsoft Azure Machine Learning and Python

✍ Scribed by Stephen F. Elston

• Publisher: O'Reilly
• Year: 2016
• Tongue: English
• Leaves: 62
• Category: Library

✦ Synopsis

Take time to explore Microsoft’s Azure machine learning platform, Azure ML—a production environment that simplifies the development and deployment of machine learning models. In this O’Reilly report, Stephen Elston from Quantia Analytics uses a complete data science example (forecasting hourly demand for a bicycle rental system) to show you how to manipulate data, construct models, and evaluate models with Azure ML.

The report walks you through key steps in the data science process from problem definition, data understanding, and feature engineering, through construction of a regression model and presentation of results. You’ll also learn how to extend Azure ML with Python. Elston uses downloadable Python code and data to demonstrate how to perform data munging, data visualization, and in-depth evaluation of model performance. At the end, you’ll learn how to publish your trained models as web services in the Azure cloud.
- With this report, you’ll learn how to:
- Navigate Azure ML Studio
- Use the Python Script module
- Load Python modules from a zip file
- Use the Sweep Parameters module
- Apply a SQL transformation
- Use the Cross Validate Model module
- Publish a scoring model as a web service to Excel
- Use Jupyter Notebooks with Azure ML

✦ Table of Contents

Copyright
Table of Contents
Chapter 1. Data Science in the Cloud with Microsoft Azure Machine Learning and Python
Introduction
Downloads
Working Between Azure ML and Spyder
Overview of Azure ML
Azure ML Studio
Getting Data In and Out of Azure ML
Modules and Datasets
Azure ML Workflows
A Regression Example
Problem and Data Overview
A First Set of Transformations
Exploring a Potential Interaction
Investigating a New Feature
A First Model
Improving the Model and Transformations
Filtering with SQLite
Another Data Transformation
Evaluating the Improved Model
Improving Model Parameter Selection in Azure ML
Cross Validation
Some Possible Next Steps
Publishing a Model as a Web Service
Using Jupyter Notebooks with Azure ML
Summary