AI / Analytics Tools in Azure

So when it comes to Artificial Intelligence in Azure, there are a lot of tools, and a lot of options and directions you can explore. And AI is a broad topic by itself. But that being said, I wanted to share some resources to help if you are looking for some demos to show the “art of the possible” or tools to start if you are a data scientist or doing this kind of work to help.

Let’s start with some demos. Here are links to some of the demos that I find particularly interesting about the capabilities provided by Azure in this space.

• Video.AI: This site allows you to upload videos and run them through a variety of cognitive / media services to showcase the capabilities.
• JFK Files: This is one of my favorites, as it shows the capabilities of cognitive search with regard to searching large datasets and making for a good reusable interface for surfacing some of the findings of things like transcription.
• Coptivity: Here’s a link to the video for CopTivity and how the use of a modern interface is interesting to law enforcement.

Now when it comes to offerings in this space, there are a lot and its always growing but I wanted to cover some at a high level that can be investigated quickly.

Cognitive Services: This includes Azure services that are more using APIs to provide AI capabilities to your applications without having to build it yourself. These include things like Custom Vision, Sentiment Analysis, and other capabilities. Here’s a video discussing it further.

DataBricks: DataBricks is a great technology for generating the compute required to run your Python, and Spark based models and does so in a way that minimizes the management demands and requirements placed on your application.

Azure Machine Learning: Specifically, this offering provides options to empower developers and data scientists to increase productivity. Here’s a video giving quick highlights of what Azure Machine Learning Studio is. And a video on data labelling in ML Studio. Here’s a video about using Azure Machine Learning Designer to democratize AI. Here’s a video on using Azure Machine Learning DataSets.

Data Studio: Along with tools like VS Code, which is a great IDE for doing Python and other work, we do provide a similar open source tool called Azure Data Studio, which can help with the data work your teams are doing. Here’s a video on how to use Jupyter notebooks with it. Additionally, VSCode provides options to support this kind of work as well (video).

Azure Cognitive Search: As I mentioned above, Search can be a great way to surface insights to your users, and here’s a video on using Cognitive Search.

Azure Data Science VM: Finally, part of the battle of doing Data Science work is maintaining all the open source tools, and leveraging them to your benefit, the amount of time required for machine configuration is not insignificant. Azure provides a VM option where you can create a VM preloaded with all the tools you need. Azure has it setup for Windows 2016, Ubuntu, CentOS. And there is even have a version built around Geo AI with ArcGIS. There is no additional charge this, as you pay for the underlying VM you are using, but Microsoft does not charge for the implementation of the data science tools on this.

I particularly love this diagram as it shows all the tools included:

Now again, this is only scratching the surface, but I think it's a powerful place to start to find out more. I have additional posts on this topic.