Streams, Lakes and Oceans - Working with Big Data with Azure ML

Explore big data architectures and Azure Machine Learning Studio integration in this comprehensive conference talk. Learn how to handle large datasets using Azure capabilities, connect to various data stores, and work with data loads, batches, and streams. Discover techniques for customizing machine learning processes with R modules, publishing endpoints for predictive analysis, and retraining models. Gain insights into scaling Azure ML web services to meet performance demands. Delve into topics such as Azure's big data capabilities, ML integration with different data stores, and designing architectures for large data volumes. By the end of this talk, acquire the knowledge to effectively use Azure ML for big data projects, adjust models using R modules, and scale solutions for optimal performance.

Syllabus

Intro
Who is Barbara
Agenda
Aerospace predictive maintenance
What is big data
Volume of big data
Traditional approach
Data ocean
Velocity
Data served
BI and analytics
Streams
ML vs Big Data
Online Learning
Big Data
Problem description
How does ML work
What I love about Cortana
Data sources
Publish as API
Architecture
Azure ML Architecture
Standard approach
Retraining
Setting up a web service
Retaining data
Scaling web services
Questions