Microsoft revamps the software's forecasting features, enabling users to fine-tune their business predictions.
Excel 2016, part of Microsoft's newly released Office 2016 productivity software suite, has new built-in tools to help users peer a little further into the future.
Data forecasting is not new to Excel. Before the new version, "many used the FORECAST() sheet function, which performs a linear forecast or extended trendlines in chart properties to extrapolate forward," reminded Microsoft's Excel team in an Office Blogs post
As software makers are wont to do, they sometimes find other ways of getting the job done with an alternative algorithm. "The new functionality in Excel 2016 utilizes another algorithm, called Exponential Smoothing or ETS," wrote the company's bloggers.
"Exponential Smoothing methods are a popular way to forecast and are among the leading methods that have become industry standards," they continued. "The main advantages of using the ETS method are the ability to detect seasonality patterns and confidence intervals."
Excel 2016 can automatically detect seasonal patterns in historical data. Microsoft recommends at least two to three full seasonal cycles for more accurate results. Users also can set a seasonality value manually.
In addition, the software's ETS-based forecasting capabilities can generate a confidence interval, which Microsoft describes as "the range surrounding each predicted value in which 95 percent of future points are expected to fall based on the forecast (with normal distribution). The confidence interval helps you figure out the accuracy of the prediction. A smaller interval implies more confidence in the prediction for the specific point."
For a guided method of creating a forecast, Microsoft suggests users click the Forecast Sheet button under the Data tab after organizing their time-based series data. "This launches the forecast dialog that walks you through the process." A how-to
is available in Excel's online documentation.
Separately, Microsoft announced new updates this week for data explorers that instead use the company's cloud-based analytics tool, Power BI.
As part of a weekly service update, the company rolled out a new Suggested Questions feature. "There are instances where you may not know the specific terms in your data (e.g. when using a dashboard created by someone else)," said Theresa Palmer, a Microsoft Power Query program manager, in an Oct. 6 announcement. "With this release, we have added a feature to help you get started with Power BI Q&A, even when you do not know anything about the data."
Power BI dashboards now display a "How to ask" link alongside the Q&A search box. After analyzing the corresponding data, the software will suggest relevant questions. Clicking a question will automatically generate a new data visualization.
Also this week, users can now insert shapes into Power BI Desktop's the report canvas, the software's authoring tool. "You can insert shapes by clicking on the 'Shapes' menu when in 'Edit Report' mode," said Palmer. "Supported shapes are rectangle, oval, line, triangle and arrow. Several formatting options are available for each shape, such as line/fill colors, rotation, title, background, etc."
Finally, Power BI dashboards can be shared without sending an email invitation. Users can now generate a copy/paste-able URL by unchecking the "Send email notification to recipients" option in the Share dialog.