In this lab we will be implementing a forecasting measure using DAX. We will not be generating a forecast data set, rather, we will create a measure that will look at the actuals from the past to forecast what will happen in the future. We will be using a CDC data set of US Deaths. In fact the Power BI Desktop file used in the lab is directly connected to the dataset, so you can simply hit Refresh in your Desktop file to get the latest data.
US deaths tend peak out in January and then then bottom out by mid-summer. This variability in values is called seasonality. You will also notice that on average we see a up trend in deaths so as the population grows so do the deaths. Therefore, in our calculation, we will need to capture both of these components, seasonality and trend.
As you will see, the expected deaths are predicted fairly well and you can also see how much higher actual deaths were in 2020 compared to what our forecast was predicting.
You read the blog and download the Power BI Desktop file used in this video here: businessintell...