In SAS Visual Forecasting, you can create a project using one of two types of data sets:
This is the source time-stamped data that you need to form time seriesan aggregation of transactional data into specified time intervals and sorted according to unique combinations of the default attributes (BY variables) vectors and to determine the best modeling nodes to use and to create forecasts. The input data set requires at least one time variable using a SAS date or datetime format. For more information, see Preparing Time Series Data.
The input data set must also include the dependent variable, which is the numeric variable that you want to forecasta numerical prediction of a future value for a specified time period for each unique combination of BY variable values. Independent variables that can explain variations in data patterns must also be included in the input data set. However, inclusion of independent variables is optional.
You also want to have several variables that define the individual time series in your data. For example, in an inventory scenario, you might want a variable that represents different regional warehouses and another variable that represents the store locations in those regions. These variables will be assigned as BY variables. They enable you to group observations into time series that can be used for hierarchical forecasting. Even more importantly, these variables will also be used as the default attributes for the project. Any other attributes that you want to include must be imported using a separate attribute data set. For more information, see Preparing Attributes for Your Project.
This is an output data set that already contains forecasts along with the historical data. Typically, this data has been exported from another forecasting project as an OUTFOR data set. External forecasts are used primarily to work with any overrides that need to be applied to the forecasts. For more information, see Working with External Forecast Projects.
The data can be imported into a new SAS Visual Forecasting project by loading it into Cloud Analytic Services (CAS). See SAS Cloud Analytic Services: User’s Guide for more information about using CAS.
Sample data sets are provided, which you can use to experiment with SAS Visual Forecasting. You can access this data at the following location: