Episode 275
| April 16, 2024
| 45 minutes
Machine Learning and Power BI
What kinds of problems are organizations solving with Machine Learning? In this episode, we explore a situation where a public works department was looking for more accurate information to predict future water levels based on rainfall to maintain water tank storage for balancing pressure and to prevent overflow flooding. Marathon data solutions consultants Brian Knox and Andy Yao, built a custom machine learning model and made the results available through Power BI reporting. We talk through some of the data hurdles the project presented, the tools they used, and how their work provided results the client could rely on. We touch on Azure ML environment and future integrations that will come with Power BI and ML.
Let us know what you think!
Have you done any work in ML or predictive modeling? Did you get any good take-aways from today's podcast? Leave us some love ❤️ on LinkedIn, Twitter/X, Facebook, or Instagram. Thank you for listening!
Our Guests
Brian Knox
Brian Knox is a senior data solutions consultant at Marathon Consulting who has passed the DA-100: Analyzing Data with Microsoft Power BI exam and has been a Power BI developer, administrator, and instructor for over seven years. He is a founder and leader of the Hampton Roads Power BI User Group. He is a graduate from the College of William and Mary, having completed the inaugural class of the Master of Science in Business Analytics program. He has an undergraduate degree in Information Science from Christopher Newport University. Brian learned early on that he wanted to combine his passions of business and technology. Brian loves all things data and is always looking for ways to share his data science knowledge to help others throughout Hampton Roads.
Andy Yao
Andy Yao is a data solutions consultant with a master's degree in Business Analytics, known for pioneering several successful machine learning initiatives. With a keen interest in analyzing complex datasets to support company strategies and enhance operational efficiencies, Andy has made meaningful contributions to the analytics community. His work, focused on applying machine learning to solve business challenges, has helped improve productivity and effectiveness across various industries. Possessing a robust foundation in predictive analytics and machine learning, Andy excels in conducting thorough research and proficiently handling technical data analytics tasks.
I think there are definitely some numbers that you can point to for a rule of thumb like the 80/20 and then scaling back if you have a smaller amount of data, but also just familiarity with your dataset goes a long way.
Meet the Hosts
Carlos Chacon
With more than 10 years of working with SQL Server, Carlos helps businesses ensure their SQL Server environments meet their users’ expectations. He can provide insights on performance, migrations, and disaster recovery. He is also active in the SQL Server community and regularly speaks at user group meetings and conferences. He helps support the free database monitoring tool found at databasehealth.com and provides training through SQL Trail events.
Eugene Meidinger
Eugene works as an independent BI consultant and Pluralsight author, specializing in Power BI and the Azure Data Platform. He has been working with data for over 8 years and speaks regularly at user groups and conferences. He also helps run the GroupBy online conference.
Kevin Feasel
Kevin is a Microsoft Data Platform MVP and proprietor of Catallaxy Services, LLC, where he specializes in T-SQL development, machine learning, and pulling rabbits out of hats on demand. He is the lead contributor to Curated SQL, president of the Triangle Area SQL Server Users Group, and author of the books PolyBase Revealed (Apress, 2020) and Finding Ghosts in Your Data: Anomaly Detection Techniques with Examples in Python (Apress, 2022). A resident of Durham, North Carolina, he can be found cycling the trails along the triangle whenever the weather's nice enough.
Want to Submit Some Feedback?
Did we miss something or not quite get it right? Want to be a guest or suggest a guest/topic for the podcast?