Machine Learning @ 1 million predictions per second and more

Watch recordings of keynote and session previews of  Microsoft Machine Learning & Data Science Summit 2016 on the latest Big Data, Machine Learning, Artificial Intelligence, and Open Source techniques and technologies.

Some take-aways from the keynote:

  1. Combination of in-memory technologies and in-database analytics with R at scale using SQL Server 2016 can make 1 million fraud predictions per second.
  2. U-SQL in combination with Cognitive APIs and Azure ML can significantly extend datasets to make possible to analyze large volumes of images (different objects and complexity) and text (subjects, key phrases, sentiments, story).
  3. In future Azure Data Lake Analytics will support Hive and Spark.
  4. Microsoft ResNet (solutions for Deep Learning) is built using 152 neural network layers.
  5. Azure N-series Virtual Machines with GPUs to be used for Deep Learning are available in preview. For example, Tesla K80 delivers 4992 CUDA cores with a dual GPU design, up to 2.91 Teraflops of double-precision and up to 8.93 Teraflops of single-precision performance.

Case Studies:

  1. Student Drop-Out Prediction Service in Indian schools uses Azure ML.
  2. PROS used Azure and R in SQL Server for airlines to recommend prices in milliseconds. For another customer they moved R-based solution to SQL Server 2016 to generate renewals automatically “faster in a factor of a hundred”.
  3. Dyxia used combination of Microsoft Band, MS Health application, Azure IoT Hub, Stream Analytics, Power BI, Machine Learning and other services to monitor and predict anxiety of children with autism.
  4. eSmart Systems created Connected Drone solution combining drones with Deep Learning in Azure to automate inspections of power lines.
  5. CrowdFlower use crowd sourcing (Human-in-the-Loop) to train machine learning models for non-confident predictions.

Below there are some screenshots from the keynote.






List of available recordings:

Vehicle Health & Driving Pattern Analysis using Cortana Analytics with Power BI

Last changes: January 13, 2016

In the following scenario of advanced analytics we will show how car dealers, insurances and automobile manufacturers can use Cortana Analytics including Power BI to gain real-time and predictive insights on vehicle health and driving pattern behavior.

The solution can be applied to following business use cases:

  • Usage-based insurance
  • Vehicle diagnostic
  • Engine emission control
  • Engine performance remapping
  • Eco-driving
  • Roadside assistance calls
  • Fleet management


Starting December 1, 2015 the solution called Vehicle Telemetry Analytics template is available at Cortana Analytics Gallery. Here is quick promotional video:

In the following video and text below you will see some details on solution architecture which includes following technologies: Event Hub, Azure Stream Analytics, Azure Machine Learning, Azure Data Factory, HDInsight, Azure Storage, Azure SQL DW, and Power BI.

Let’s look on data flow and solution components.


The Event Hub is used to ingest huge amount of events from the vehicles into Azure for real-time and batch analytics.

The Stream Analytics job is performing real-time data ingestion into the long term storage for batch analytics and data preparation for real-time predictive insights.

Below you can see description of three queries processed in the Stream Analytics for following purposes. (All three queries are enriched with detailed data on each vehicle from Blob Storage).

Query #1 performs join with reference data from Azure Blob Storage and accumulates the resultant data into a different container in the Blob Storage for rich batch analytics.

Query #2 publishes the data as-is to the output Event Hub so that it can be consumed by the RealtimeDashboard app that invokes machine learning request/response end-point for real-time anomaly detection and pushes the results to the PowerBI live dashboard.

Query #3 performs aggregations on the data within a 3 sec tumbling window and publishes it to an Azure SQL instance that got provisioned as part of the deployment.

Data Factory is used for

  • Orchestration, monitoring and management of the batch analytics pipeline
  • Transformation of the data in an on-demand HDInisght cluster for rich insights on Driving Behavior Pattern and Vehicle Health Trending
  • Data movement across the various data stores


All data in source datasets are processed using Hive queries where we describe data structures based on CSV files. Additionally we define new tables and calculate aggregations using INSERT request.


In this solution, we are targeting the following batch insights:

  • Aggressive driving behavior (Identifies the trend of the models, locations, driving conditions, and time of the year to gain insights on aggressive driving pattern allowing Contoso Motors to use it for marketing campaigns, driving new personalized features and usage based insurance.)
  • Fuel efficient driving behavior (Identifies the trend of the models, locations, driving conditions, and time of the year to gain insights on fuel efficient driving pattern allowing Contoso Motors to use it for marketing campaigns, driving new features and proactive reporting to the drivers for cost effective and environment friendly driving habits.)
  • Recall models (Identifies models requiring recalls by anomaly detection trend and correlation with driving habits)

An anomaly detection Azure Machine Learning model is used in this demo to detect safety issues for vehicle recall and identifying vehicles requiring maintenance. This model is published in an existing subscription and the web service endpoint is leveraged both in request/response and batch mode for operationalization in the real-time and the batch processing.


Aggregated data from Blob Storage is moved to Azure Data Warehouse for historical storage.

Power BI dashboards contain historical data from Azure DW and real-time data from the Azure Stream Analytics and the Event Hub.


Special thanks to authors of the demo scenario: Anand Subbaraj, Sanjay Soni, Christoph Schuler, Santosh Waghmare, Shashank Khedikar, and Sam Istephan.


Expense Management Scenarios


The Microsoft Analytics Platform System

This video introduces the Microsoft Analytics Platform System which brings together a high performance MPP RDBMS (PDW) with Hadoop, seamlessly integrating data of all sizes and types, offering the perfect platform for a company’s modern data warehousing needs.


Analysis of Big Data for Financial Services Institutions

In this blog post, we will look at analysis of stock prices and dividends by industry. This task is important to all participants of Stock Market including individual retail investors, institutional investors such as mutual funds, banks, insurance companies and hedge funds, and publicly traded corporations trading in their own shares.

In this demo, team of Stock Trading Company analyses semi-structured stock data from the New York Stock Exchange (NYSE).

  1. Data Architect collects data and makes information accessible to business. He will use Hadoop-based distribution on Windows Azure and Hive queries to aggregate stock and dividend data by years.
  2. Financial Analyst will analyze stock data and prepare ad-hoc reports to support trading and management processes. She will use Power Query add-in for Excel to join aggregated data from Hadoop with additional information on top 500 S&P companies from Azure Marketplace Datamarket. Additionally she will create ad-hoc reports with Power View for Excel.
  3. Trading Executive is responsible for understanding key decision makers and suggesting best product mix of securities. He will make some modifications to Power View reports provided by Financial Analyst.

Details on how Data Architect aggregates data in Hadoop are available in a separate blog post.

Below you can see some screenshots from the demo.










Aggregating Big Data with HDInsight (Hadoop) on Azure

When we a talking about Big Data we may mean huge amounts of data (high Volume), data in any format (high Variety), and streaming data (appearing with high Velocity). Microsoft provides solutions for all of these “3V” tasks under unified monitoring, management and security, as well as unified data movement technologies. These
workloads are supported correspondingly by SQL Server Database and Parallel
Data Warehouse, HDInsight (Hadoop for Windows or Azure), and Microsoft SQL
Server StreamInsight.


Let us talk about Microsoft Big Data technology for Non-Relational data.

Microsoft’s adaptation of Hadoop technology can be deployed in a cloud-based environment or on-premises. The Hadoop-based service on the Windows Azure platform is a cloud-based service that offers elastic (in a term of data volumes) analytics on Microsoft’s cloud platform. For customers who want to keep the data within their data centers, Microsoft provides Hadoop-based distribution on Windows Server.

In this blog post, we will start diving into Hadoop in Azure technology and Hive queries to analyze semi-structured data in Hadoop.

In addition to traditional data warehousing, when operational data stored in special structures in Enterprise Data Warehouse, we can store all other raw data in “Store it All” cluster. At any moment, we are able to create query to these data to answer some business question. (In addition, we may store the answer in the Data Warehouse if necessary)


Let me introduce the first part of Bid Data Demonstration where Data Architect will store log files with stock prices and dividends in Azure Blob Storage and will use Hive queries to aggregate data by years and stock tickers into separate file.


Here is the video:

Additional materials: Windows Azure Storage Architecture Overview

Oil and Gas Analytics with Power View for SharePoint

Oil and Gas industry requires highest quality of decision-making based on correct and intuitive data. The industry becomes more innovative and requires faster usage of new production methods paying attention to eliminate risks at the same time.

Microsoft Business Intelligence provides tools to analyze historical data in intuitive and interactive way.

In this demo, team of Oil and Gas Company solves business issue through collaboration of following persons:

  1. Chief Executive Officer develops strategy, monitors operating efficiency, and achieves growth of profit and company value.
  2. Business Unit Manager works to ensure business continuity, improve technology, and drive excellence.
  3. HR Executive is responsible for improvement of personnel performance through training and appraisals.

All of them are using the same company portal built on SharePoint portal with a set of pages containing interactive reports (dashboards). These reports based on Business Intelligence Semantic Model (BISM) and visualized with Power View for SharePoint.