[I’m Speaking!]: PASS DevOps Virtual Group

As the use of Machine Learning becomes more integrated into our Analytics core competencies, I find myself reviewing implementations and processes and asking myself … ‘How do I make this easier for next time’? CI/CD practices in Machine Learning projects is a step that is often overlooked in newly formed teams.

I am excited to be speaking with the DevOps enthusiasts @ PASS DEVOPS VG about how they can help their Data Scientists operationalize their code. Register: DEVOPS.PASS.ORG
Date/Time: Wed, May 22 7:00 PM – 9:00 PM GMT

Azure offers the right tools and services that make the process seamless. During the session, we will walk you through a simple process of integrationg Python code in VSCode with Azure Devops capabilities. Oh, and spoiler alert, as an #AIAdvocate, I might also show you a bit of what #Azure #CognitiveServices can do!

AI in Power BI: Building the Customer Demo that Checks All the Boxes, and more!

I recently responded to a common ask. An existing customer is evaluating reporting tools, do you want to show them what Power BI can do? Of course, it was on!

My first ask: Get me some randomized sample data. This is key to speaking to your customer needs and wants. Be relevant! What is a good dataset? Financial data that covers more than two years, with hierarchical references is perfect. Location hierarchy, such as Country, State, City or category hierarchy such as Region, Department will work. Be sure to include some type of financial transaction field, as well as a transaction datetime timestamp. Hierarchical data will allow for drill downs and the ability to show both summarized and transaction level data.

Now it comes down to building the report and dashboard, and less is absolutely more! What features should we add to the demo report and dashboard?

Speak to Your User Base

Power BI is an enterprise level reporting tool that has the security controls and extensible framework that when properly deployed can be leveraged for both internal Enterprise level reporting; as well as servicing assets consumed externally by 3rd parties.

Plan on showcasing Multi-Tenancy features by configuring row-level security on your data set. Add a few user roles and configure a related group of visuals to change based on user role, including a data table, a map and image. Cover the basic visualizations, slicing and filtering capabilities. The conversation will naturally speak to the white labeling and re-branding capabilities of Power BI, and you might find that the potential user group or use cases within the customer organization will increase dramatically.

Recruit a developer to put together a quick splash page similar to their existing reporting interface to embed the Power BI report. New technologies can be overwhelming, and it can be tough for anyone to visualize how a feature will function within their existing eco-system. If you can get a sample of their data to build your demo, the audience will already be familiar with the data and you relieve a lot of the “imagining” that your audience would otherwise have to do during your demo.

How easy is Power BI to learn and use?

At this point, you’ve successfully shown Power BI’s ability to serve reports and we can now show how easy it is to author and manage reports. A few minutes on creating a new dataset, adding a new visualization and subscribing to a report covers how quickly users can connect to data sources, author reports and self-service. This is also a good time to introduce the marketplace and custom visuals, as well as the .NET API. Even the pickiest developer in the room will agree that while the interface is user friendly, the capabilities for extensive customization is also available.

Looking forward. Why invest in Power BI?

Now it’s time to show them things they weren’t expecting! Your customer is probably not aware of some of the newer artificial intelligence features available within Power BI. Users can now integrate directly with machine learning models for text and image processing while also leveraging Microsoft Natural Language Processing (NLP) features. Key Influencers allows users to perform basket analysis utilizing fields within their dataset. Q&A capabilities allow users to ask questions of their data. Data scientists can easily add in and tweak Python & R code snippets.

At this point, you’ve shared the sheer velocity of development and availability of Microsoft AI features in Power BI.  Congratulations!  #AIforEveryone

Key Influencers Visualization: Courtesy of docs.microsoft.com

Global AI Bootcamp – Houston Edition 2018

Thankful. Humbled. Grateful. [I’m Hosting!] I am honored to be hosting Global AI Bootcamp-Houston Edition, part of a global event sponsored by the Microsoft AI School. Register Here: https://lnkd.in/e9gV9bh . More details soon!
Thank you to our local sponsor: Avanade, and special thanks to Kyle Hale and my sqlfamily for your support and mentorship. Looking forward to working with hashtagGlobalAISummit HQ, Henk Boelman and Alexander Meijers.