Machine Learning Without the Complexity

By | 2015-08-10T13:37:36+00:00 August 10th, 2015|Azure Machine Learning (AML), Cortana Analytics|0 Comments

Northwest Cadence is excited about the power Cortana Analytics brings to businesses big and small. I am too! But, when I was asked to take the lead with our Cortana Analytics offerings, I didn’t know where to begin. The big data and advanced analytics space can be a bit overwhelming. After all, that’s what data scientists are for! While my background is in mathematics as opposed to data science, I remember taking a machine learning class in college and found it to be very interesting. As Northwest Cadence ramped up our Azure Cortana Analytics business, Azure Machine Learning (AML) seemed like a good place to start.

For the past year, Northwest Cadence has been building out a data analytics platform to gain insights from software development data. As I created an AML workspace, I remembered how complex machine learning was in college. The class was full of statistics, graph theory, and complicated implementations of neural networks and the like. I was not looking forward to going through all of that again; and as it turned out, I didn’t have to! The UI was a pleasant surprise. AML models are created through a graphical connection of prewritten modules. All of the machine learning models you’ll need are already included and can be adjusted as needed. These are the same models Microsoft uses for anomaly detection, regression, clustering, and classification. In fact, if you’ve ever searched for something with Bing, you’ve already used them.

Microsoft Azure Machine Learning Studio

In no time at all I was creating fun and insightful experiments with AML. Just drag in the modules you want and wire them up. It even has built-in preprocessing capabilities, so you can easily import your data from on-premises or the cloud and get right to work. If you need to use custom Python or R code, no problem, there are modules to handle that too. The best part is, once you’re done you can easily publish your model and use it via a REST API call, then visualize the results with Power BI! Very cool.

If you haven’t already, take a few minutes to explore AML. There are usable datasets included in the workspace so you can play around and casually build insightful models without a bunch of aggregated data of your own or even intimate knowledge of the mechanics of machine learning. Microsoft has really been working hard to make Cortana Analytics easy to use for those of us who aren’t data scientists without sacrificing serious functionality and results.

About the Author:

Leave A Comment