I’ve always said if you call PowerBI a visualization tool, you don’t entirely know its capabilities. Also, for entry-level analysts, the analysis goes beyond exploratory analysis. You can take your analysis one level higher with anomaly detection and ask why it is happening.
Good news, you don’t even need to learn any new thing.
So, what is anomaly detection? Anomaly detection is simply identifying behaviour or patterns that differ from the norm. In terms of data, it merely recognizes the data points that are different from the normal behaviour of all other data points.
How, then, do we do this in PowerBI? Let’s see how it is not complicated at all.
I’m using the popular SuperStore data, and the model I’ve created is the one you’ve made if you’ve worked with this data. Nothing complicated, as you can see.
Measures used include the following;
As you can see, the measures are nothing complicated; they are things you are used to—pretty straightforward, right? I’ve also created a set of charts using the model and DAX measures above.
The first chart is a scatter plot comparing each city’s total sales and year-on-year sales differences.
The bar chart shows the current sales and sales LY by Product, while the line chart shows cumulative totals.
Let’s focus on Q2 2017. We are comparing the Year on year sales with Q2 2016
Here we drill down to the dates.
Let’s compare performance in Q4–2016 with Q4–2015
You can see with just a simple model coupled with simple DAX, you can do more than just saying so, so made the highest/lowest sales. You don’t need to understand complicated dax to take your analysis to the next level.
I hope you can now see PowerBI as more than just another visualization tool.