In the September 2016 update of the Power BI Desktop, the Power BI team took the first step toward providing Power BI users with the possibility to add forecasting to charts. Forecasting, a fundamental component of data mining, is one of the first steps towards providing artificial intelligence to users of Power BI. Imagine adding forecasting to your reports to help you see when you will reach your business goals. On our team in BI@Microsoft, we are using this feature to predict when we will reach our Monthly Active User (MAU) goal for Power BI usage.
You can get an excellent sense of how you can use forecasting in this 6-minute video.
When I tell people about this great new feature, they often ask, "What forecasting algorithm is Power BI using?" Power BI documentation will soon include this information, but in the meantime, here's the answer. The Power BI Desktop uses an assembly forecaster. Basically, Power BI picks the algorithm it thinks best fits the data. As of this update, the assembly forecaster includes exponential smoothing.
The Power BI team plans to continue to improve their forecasting algorithms in future releases of the Desktop. Ultimately, the team also plans to add forecasting to the entire Power BI suite of tools. Soon you will be able to have forecasting in the reports and dashboards you give your users in the Power BI service. For now, get started by adding the feature to your existing reports in the Desktop so that when forecasting is available in the Power BI service, you will have your reports all ready to publish and distribute to your users.
Note that forecasting is supported only in the Power BI Desktop, and not yet in the Power BI service. Also, for now you can only add forecasting to a single value line chart.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.