Back to List

I’ll Have an Azure Christmas

Dale Mumper & Jeff Hallett  |  
Dec 21, 2015

It is a sunny breezy day on one of the 700 islands that make up the Bahamas. Santa and Mrs. Claus are relaxing and enjoying the last days of the off season. And you thought Santa’s nose was red from frost bite? No, that is only half the story. Just as Santa was about to fall asleep for an afternoon nap, his cell phone rang. He quickly fumbled to stop the “I Saw Mommy Kissing Santa Claus” ringtone. It is Mitch Winter on the other end of the line, the head elf back at Santa’s workshop. He is excited and talking rather fast, and has some ideas to improve the coming holiday season.

Santa’s head is barely raised off his beach chair as Mitch begins to explain. He had attended a webinar with his Microsoft rep and says there are some great new technologies that he thinks can make this Christmas better. Specifically, Mitch wants to address the need to substitute gifts depending on whether kids are naughty or nice in the days right before Christmas. Additionally, Mitch expresses some frustration with all the toy possibilities and how the toys and gadget choices are so plentiful that they could use some advice on how many of each of the new gadgets he should build. In short, he needs some help predicting demand.

Microsoft has some tools in Azure and Mitch thinks that can help with both of these scenarios. Santa is really not in the mood for shop talk but he knows that an excited elf is a good thing and he does not want to discourage this. Nothing makes an elf happier than being efficient. Santa refers to this as “Elf-iciency”. Santa agrees to let Mitch give it a try since Mitch says he has everything he needs to do this in 30 days and it won’t cost Santa a dime to try it with a free trial on the Azure platform.

The workshop is abuzz as Mr. and Mrs. Claus enter. Of course they are tan and feeling a bit weary from their travels home but they can tell something exciting is going on. John Kringle, the head of IT at the workshop, is giving a presentation about a new dashboard he had developed and has the complete attention of all the Elves.

John explains how this is a real-time dashboard that is continuously fed new information based on the input from individual parents. They have written an app that is deployed on all the major phone platforms. This data is streamed from the apps to an Azure Event Hub in real-time. The data can be displayed in real-time on a Power BI dashboard. This will allow for production in the workshop to be more elf-icent because they have better insight into the demand and can tailor the labor and materials to be exactly what Santa will need. This will be a great boost to workshop operations.

Mitch then takes the stage and displays another dashboard. The busy elves have gathered demographic data and purchased some retail sales data from the Azure Marketplace. This will give them much better insight into what is trending with each group of children. This will allow the workshop to better satisfy kids demand and predict gifts more accurately.

Mitch speaks to some data quality issues they have had in the past. Some of the children had self-reported their naughty or nice classifications, and others were putting down inappropriate gifts for their siblings as a trick. Luckily, with the Azure Marketplace and Azure Machine Learning, we have access to machine learning algorithms that can help us be more accurate in our gifts. For example, businesses have long had the need for fraud detection software, more loosely known as anomaly detection. We can use this to tell us if children actually are naughty or nice. (Note that Mitch’s brother-in-law had previously used this type of technology to determine if prospects on a customer list would actually turn into buyers or not). We add in data factors that we have accumulated (from open sources, not the NSA) and can determine with 95% accuracy which category a child belongs to. What about that other 5% that are misclassified? Well, this is a proof of concept after all. Just consider that reinforcement for even better behavior next year!

For picking gifts for children, things are nowhere near as easy as they once were. Remember the days where Johnny wants a pair of skates, Susy wants a dolly, and Nellie wants a story book? Those times are long gone. Now there are bits and bytes and e-this and e-that, things requiring specialized skillsets and technologies here in Santa’s Toy Shop. Hoverboards? Good heavens! So what we’re able to do now is grab a data extract from toy manufacturers indicating their product specifications and target demographics. We import that into our Azure HDInsight with Apache Spark cluster to do some heavy data analysis, with each child on the planet getting some attention. After pushing a button, datasets are sent to Machine Learning algorithms to help identify the perfect group of items for that customer, er, I mean child. It really simplifies the process now that we have millions of toy SKU’s and billions of children.

Santa is good at this, but now we have data!

Mitch had saved the best for last.  In each toy the elves are able to embed small IOT transmitters. There are no privacy concerns since they just report how much each toy is played with through motion sensors. It gives real-time data again via Azure and Power BI. These elves will love seeing the instant feedback of how these toys have been accepted. But the real value Mitch explains is that this usage data will help them better understand if the toy breaks and how long it is used. This is great to make the toys and gadgets even better as well as have quantifiable support indicating how well liked they were.

After listening to all of this, Santa realized that Mitch and the elves had designed the perfect Christmas present fulfillment system. By using cloud services that they only pay for by consumption, they’re able to avoid depreciating a datacenter full of aging servers. They can focus on building toys and delivering joy instead of patching servers and replacing hardware. By using new technologies like the Internet of Things and Azure Event Hubs, the massive scale of Santa data can flow without concern. And once all of the data is safely accumulated in their Azure Blob stores (Santa still doesn’t quite get that one), John Kringle and the rest of the Analytic Elves can use machine learning to make the right decisions at the right times. It’s looking to be a wonderful season and Santa can’t wait to meet this new elf called Azure.



Love our Blogs?

Sign up to get notified of new Skyline posts.


Related Content

Blog Article
Azure Backup Simplifies Your Cloud Backup and Recovery Strategy
Erik DempseyErik Dempsey  |  
Feb 27, 2018
We recently started implementing Azure Backup here at Skyline, and I was tasked with digging into the service to find out capabilities, costs, retention, etc. I have to say, after all the research, lab work, testing, and move to production, I could not be happier with all the different options...
Blog Article
Power BI: Drillthrough Using Built-in Date Hierarchy
Scott HietpasScott Hietpas  |  
Jan 16, 2018
The January 2018 Power BI Desktop update introduced a preview feature that allows you to see the auto-generated date hierarchy in the data model. Per the January 2018 feature summary…For a long time now, we’ve been automatically creating a date hierarchy for you whenever you are...
Blog Article
Power BI & Excel - Options for Data Sharing and Publishing
Kathy ShieldsKathy Shields  |  
Nov 14, 2017
If you’re a Power BI user or support those who are, you know Microsoft continues to add features for data sharing and report publishing between Excel and Power BI. Many Power BI users are also strong Excel users, relying on it for complex analysis and reporting on data.  There are...
Blog Article
How to Quickly Detect Tabular Referential Integrity Issues
Cory CundyCory Cundy  |  
Nov 07, 2017
Download Tabular RI Analyzer Often when we think about referential integrity (RI) we think of relational databases or data warehouses. While we should have referential integrity constraints in those systems, it may not always be implemented for various reasons. When building a tabular model, we...
Blog Article
The Business Value of Data Visualization
Josh PinsonneaultJosh Pinsonneault  |  
Sep 21, 2017
Data is abundant everywhere in our lives whether we’re at work, school, or at home. However, data “in the wild” is often raw and insights can be lost without visualization. When faced with many different data points, it can be difficult for our minds to effectively process and...