Archive for March, 2016

Deploy from RStudio to Azure

March 6, 2016

Azure offers several tools that can help Data Scientists manage and publish their predictive models. These tools greatly shorten the task of making the predictive model consumable by either users or applications. One such tool is the AzureML R package located on CRAN (The Comprehensive R Archive Network).

The AzureML package allows a Data Scientist to quickly and easily publish a predictive model to Azure. Azure will then create a Web Service (HTTP REST API) for application/user query and response solutions. Additionally, Azure also provides an Azure ML Request-Response Service Web App template that can be configured and deployed in a matter of minutes.

deployRStudioAzure

Using this simple deployment model a predictive model can go from local (PC) to global (Azure hosted Web Application) in a matter of minutes.

The IrisPublishWebService.R RScript will download the Iris dataset, create a predictive model and then publish it to Azure as a HTTP REST API Web Service. A Web Application can then be used to consume the web service. See final solution IrisClassDemo.

All this can be created in a matter of minutes using Azure!

Setp-by-Step YouTube Video: Deploy from RStudio to Azure

Advertisements