In the September update of Power BI, Microsoft released the preview version of a forecasting feature for a line chart.
Before we dive into Forecasting in Power BI, let’s quickly cover some points about around forecasting:
What is Forecasting? Simply put, is predicting or estimating (a future event or trend).
Seasonality: In any time series data, if we have any yearly predictable change or pattern then we have seasonality in the data set. In Power BI we can indicate seasonality in our data, 12 for yearly, 6 for half yearly and 4 for quarterly seasonality.
Confidence interval: It’s a probability defined in such way that actual values will lies with in this range, for example if we provide a 95% confidence interval, then we are saying that 95% is the probability of the actual value lying within the range.
Ignore Last: Certain data sets can have incomplete data for the last x months. We can that the last 1 or more months has incomplete data and should be ignored for prediction purposes.
Forecasting in Power BI
Data set requirement: To use Power BI forecasting, we need to have time series data, therefore we need a date column in the data set. For our example, I will be using a basic ticket counting data set. It consists of the count of ticket creation over the last 3 years.
Here’s the file: