Operationalizing R Models on the Azure Cloud

Speaker: Robert M. Horton, PhD MS, Senior Data Scientist, Microsoft

I will begin with a brief discussion of cloud computing, with examples relevant to R users. These include:

• Infrastructure as a Service (running virtual machines on the cloud)

• Platform as a Service (starting and using an AzureSQL database)

• Software as a Service (an overview of Azure Machine Learning)

We will use these tools to train some simple machine learning models and deploy them as predictive analytics services on the Microsoft Azure cloud using the AzureML R package, then call these services from Excel and R. Along the way we will explore the structures of several classes of R model, and write wrapper functions that use these models to predict outcomes from new data.

Bob got his PhD in Biomedical Sciences at the Mayo Clinic in 1991, where he studied structure-function relationships in genes of the mouse major histocompatibility complex and helped to develop and popularize some widely used genetic engineering techniques. As the field became increasingly computerized, he moved to Internet evangelism (editing the "Internet OnRamp" column in "Biotechniques" from 1996-2000), and helped develop the infrastructure for the Dermatology OnLine Journal, the first all-electronic medical journal in Medline. He finally went all digital and got a degree in computer science (CSU Sacramento, 2005) and started a company to develop simulation-based educational software for molecular biology (cybertory.org, cybertory.com). He is an avid R programmer, and has worked with multiple Silicon Valley startups. He currently holds an adjunct faculty position in Health Informatics at the University of San Francisco, and is a Senior Data Scientist at Microsoft.
6:30 Doors open - socializing & individual questions

7:00 Announcements & Introduction

7:10 Speaker(s)

8:30 Wrap up & socializing

9:00 Out the door!