Use the Input Data panel of the Predictor wizard to select the location and arrangement of historical data to analyze.
Tip: | After you start Predictor the first time, Input Data opens automatically whenever you start Predictor, or click Input Data in the navigation pane of the Predictor wizard. |
To select the location and arrangement of historical data:
Open a model with historical data, select a data cell in the range to analyze, and start Predictor as described in Starting Predictor and Running a Forecast.
Input Data shows a possible data selection in the Location of data series text box and the illustration at the right side of the panel.
Location of data series indicates the cells that contain data to analyze. If the data series have headers or labels at the beginning of the rows or columns of data, include them in the selection and select the appropriate Headers settings. If necessary, select a different data range.
Note: | If you select one cell before you start the wizard, the data range is selected automatically, based on the continuously filled cells around the selected cell. If you select a range of cells before you start the wizard, that range is selected. If you do not select a cell, or if you select an empty cell before you start the wizard, you can select the range using the cell selector. You can have discontiguous data series with blank columns or rows between them. For selection rules, see Selecting Discontiguous Data. |
Confirm that the Orientation, Headers, and Labels settings are correct:
Orientation—Specifies whether data series are in rows or columns: Data in rows indicates that historical data is in horizontal rows; Data in columns indicates that historical data is in vertical columns.
First row (or column) has headers—Indicates whether the selected data has a title or header cell at the top of each column (if the data is in columns) or to the left of each row (if the data is in rows).
First column (or row) has dates—Indicates whether the data range has a first row or column for dates. Predictor recognizes dates only in cells that are formatted as Microsoft Excel dates.
Back—Opens the Welcome panel
Next—Opens the Data Attributes panel
Run—Runs Predictor if all required settings are complete, using the current method selections
Close—Closes the Predictor wizard
Help—Displays online help for the current panel
When settings are complete, click Next to open Data Attributes and set seasonality and optional events and screening options. For instructions, see Selecting Data Attributes—Seasonality, Events, Screening.
Note: | If the data range has empty cells in the middle of a data series, by default Predictor fills in the missing data (see Viewing Screened Data). If you select multiple data series, the data series are not required to start at the same time period. However, all the data series must end at the same time period. |
Tip: | For a quick forecast, complete the Input Data settings and click Run. Logical defaults on the remaining panels help ensure accurate results after you select a range of historical data to analyze. |