In this post, I want to dive into anomalies and outliers in your data and how you can analyze and showcase them effectively inside of Power BI.
What I will do is give you a quick explanation of outliers and anomalies. Then leave the rest of the information to the embedded videos to showcase the techniques in greater detail.
Anomalies in your data are results you don't expect or that surprise. Therefore, through your analysis, you will likely want to segment these specific results in your data.
To do this in Power BI, you will need to combine a number of DAX formula and visualization techniques to showcase what is an abnormal result and why this anomaly may be occurring.
This is what this tutorial is all about - how you can apply and implement these techniques successfully in your own reports.
Outliers are very similar. They are results you wouldn't expect based on historical averages or results.
But I do classify them slightly differently to anomalies because you may want to put trigger points around what you would consider an outlier. The trigger point is likely a calculation level, around which you may be looking to review results that are continuously above (or below).
This is what I will go into with this outlier detection tutorial. I will show you how you can discover these outliers or at least showcase these really effectively in a compelling way using charts inside of Power BI.
There are a substantial number of great Power BI features and techniques I run through in this video.
There is also a workshop I recently completed about outliers - Detecting and Showcasing Outliers. This was attended by around 200 people live and it was a very successful event.
If you have the time and want to review how you can build an entire reporting model around these outliers, then this is a must watch webinar and workshop. There are details in this workshop that will enable you to truly comprehend what is required to drill and dive into outlier results.
You will be able to understand why they are occurring and what's causing these outliers overtime. There's plenty to review in this detailed workshop.
Hopefully, this gives you a really good idea of what anomalies and outliers are and how you can successfully implement this logic in the development work you're doing inside of Power BI.
I feel that if you can implement this logic, it will be considered valuable work by your consumers. This could be the difference between good results and poor results, good decisions versus bad decisions. It is absolutely important to showcase this well on a consistent basis inside of Power BI.