The September 2016 release of Power BI desktop marks a preview of the forecasting feature. In this post, you’ll see how you can enable it and use it. Also I’ll share some perspective on how to use forecasting to make better business decisions. Let’s start with a video showing how to use the new forecasting feature.
This Harvard Business Review article How to choose the right forecasting technique (may require registration) offers a really nice summary of techniques you can apply and the best options for using them. Simply put, it’s not that easy to get it right. Seems like everyone, the Small Business Administration, Investopedia, Inc, iSixSigma, Business Insider, and surprisingly even WikiHow have solid content offering insight into the art and science of building business forecasts. Of course some formal education on the topic might be of interest before you stake you business on the resulting forecast.
While a built-in UX feature is great for forecasting, sometimes you want more control. To that end, Power BI supports R, where you can do customized forecasting using a number of different techniques. The forecast package can be obtained from CRAN. The Power BI R Script Showcase contains a forecast example to help you get started. While R is more technical and requires a bigger investment of your time to learn, you might find that using R to really understand the principles of forecasting will enable you to make better forecasts and crucially understand the limitations of what a forecast can and cannot tell you.
Last but not least, now that you can pin tiles from Excel files to your Power BI dashboards, you might also want to explore Excel 2016’s built-in forecasting options that are backed by Excel functions (Forecast.ETS, Forecast.ETS.SEASONALITY, Forecast.LINEAR, Forecast.ETS.CONFINT, Forecast.ETS.STAT). Again, you get some UI to help you get started (it might be more approachable than R) and you get some functions that help you understand the details of the forecast.
Power BI’s forecasting feature is starting it’s preview. Give it a try, but as with any forecast double check your assumptions before you proceed.
- Data Made Good