What: Machine Learning Week, sponsored by Predictive Analytics World. Where: This is a virtual event. When: Sunday, May 31st through Thursday, June 4th. Registration is $990 for the livestream and $1490 for livestream and recordings. Register on the Predictive Analytics World website.
What I’m Presenting
Wednesday, 2:45 PM — 3:05 PM PDT — Forecasting Demand in the e-Commerce Space
This is the first time I’m giving a public presentation explicitly related to what my company does. Typically, I present on tangential things: database administration, data science, security. But here, it’s front-and-center. This made it an interesting challenge to provide useful information without laying out any proprietary company information
What: SQL Saturday Brisbane. Where: Internet Australia. You have to turn your monitor upside-down to see it correctly. When: Saturday, May 30th. Admission is free. Register on the SQL Saturday website.
Please note that all of the times are in Australian Eastern Standard Time (AEST). This is UTC+10, so if you’re on the Eastern Daylight Time zone like me (UTC-4), that’s 14 hours ahead. In other words, I present on Friday starting at 10:15 PM EDT.
I’m presenting a full-day training at PASS Summit again this year. Here are the details:
SQL Server on the Edge: IoT with SQL Server and .NET Core
In this day-long training session, you will learn about Azure SQL Database Edge, the version of SQL Server intended to run on Internet of Things (IoT) devices. We will discuss the types of scenarios we might try to solve using IoT devices. From there, we will learn about device management through the Azure IoT Hub, including installation of applications from the Azure marketplace, as well as the development and deployment of custom Docker containers in an Azure Container Registry. Over the course of the day, we will build out a practical scenario and take a look at features in Azure SQL Database Edge, including handling time series data and machine learning on edge devices. As we expand out from a single device, we will learn how to automate deployment and updates at scale using capabilities in Azure IoT Hub.
Prerequisites: A solid knowledge of T-SQL, a basic understanding of Docker and containers, some knowledge of .NET (C# or F#).
What Can You Expect?
I’m still working out all of the details, but here are the top-level items:
Module 0 — Prep Work
In the first module, we will understand why we might want to care about the Internet of Things, looking at the types of scenarios IoT can solve.
Module 1 — Configuring Azure IoT Edge Devices
In this module, we will set up Azure IoT hub, take a look at the Raspberry Pi 4, and install Azure IoT Edge on the Pi. We will also see how to connect a virtual machine to Azure IoT Hub to assist with testing.
Module 2 — Azure SQL Database Edge Installation
Once we have a device in Azure IoT Hub, we will see how to install Azure SQL Database Edge, including configuration and deployment of dacpacs on a VM and on a Raspberry Pi 4.
Module 3 — Developing and Deploying an Application
With a database in place, we will work on an IoT solution in .NET Core and connect to our Azure SQL Database Edge instance.
Module 4 — Diving into Time Series
One of the main promises of Azure SQL Database Edge is the support for time series, and we will investigate what is available on that front.
Module 5 — Machine Learning on the Edge
In this module, we will review ONNX, the Open Neural Network Exchange. We will see how to train a model on a host, deploy it to Azure SQL Database Edge, and predict using the native PREDICT operator.
Module 6 — Device Management
The final module will extend us beyond a single device, as scale is the name of the game with IoT. We will also look at tools available for monitoring and providing insights.
Course Objectives
Upon completion of this course, attendees will be able to:
Configure an Azure IoT Hub
Connect IoT edge devices (such as the Raspberry Pi) to Azure IoT Hub
Deploy Azure SQL Database Edge to edge devices en masse
Develop and deploy custom .NET code using containers
Well, it’s kind of an old video. You can tell because I could still get away with going to Hope, Washington at the time.
If John Rambo wants to go through your town, just let him.
A version of this video was first released for the PASS Insights newsletter, and I’ve held off for a bit before posting this version to let it ride its course.
If you want to get the code used in the demo, grab it from my GitHub repo.
I’ve had on my agenda to do more of these videos, so hopefully I’ll get a cadence going, and that someday it will be a regular cadence. That’d be great.
Data Architecture Day was this past Saturday, and it was a hoot. In case you missed it, all of the sessions are up on a YouTube playlist.
Wait…How Many Sessions?
Data Architecture Day was quite a full day, with 11 sessions over 13 hours, with short breaks in between sessions.
The only unfortunate part about doing this as a marathon is that I had to be there for each session as MC and producer, so I had to be on for 13 hours. But that’s still easier than presenting!
Speaking of presenting, thank you to each presenter who took the time out of their schedules to share knowledge with us. They are, in order: Neil Hambly, Andy Leonard, Philipp Krenn, Joe Karlsson, Johan Ludwig Brattås, Louis Davidson, Solomon Rutzky, Matt Gordon, Angela Henry, Hasan Savran, and Ginger Grant.
Quick Stats
Throughout the day, we had 376 unique viewers, including a mid-day raid:
Without the raid, we had a fairly consistent behavior, with one stretch from 9 AM until about 4 PM, and then a smaller but still reasonably-sized group for the second half of the day. The big drop-off around 12:15 PM is when we went to a 20-minute lunch break.
We had 935 live views throughout the day, which says that people would leave for a while and come back. Not surprisingly, the bursts of people entering typically happened at the beginning of each session.
Two Big Questions
With Data Architecture Day 2020 in the books, there are two questions left.
The first question is, will there be a Data Architecture Day 2021? That’s a good question, and I don’t know the answer yet. We put this event together in response to the necessary cancellation of so many data platform conferences. But as conferences (probably) return in 2021, we will have to see if it makes sense to put together another Data Architecture Day next year.
The second question is, will there be other TriPASS-sponsored online conferences? The answer to that is a resounding probably. During this week’s episode of Shop Talk, we asked viewers for ideas on topics. Whatever topic we choose, we’ll limit it to 5-6 sessions max as I have learned my limits. But if you have great ideas for a mini-conference, let me know and we’ll pick an idea and (probably) organize an event for later in the summer.
What: SQL Saturday Richmond. Where: On your computer. It’s a virtual event. When: Saturday, April 25th. Admission is free. Register on the SQL Saturday website.
This is the first ever virtual SQL Saturday, so I’m glad I could take part in it. Although it’d be nicer if I could be in Richmond and present in person, I’m happy to have the opportunity at least to present from afar.
