Selecting data sources

  • Spatial temporal prediction requires a context data source.
  • The context data source identifies features on the map. The context data can also contain fields that can be used as inputs for the model.
  • The context data source contains separate entries for each location for each time period, and the data source must include one or more fields that define time periods.
  • You can also include a point density data source. A point density data source contains coordinates, such as longitude and latitude) and one or more fields that define time periods.
  • If you include a point density data source, the calculated kernel density estimate (KDE) is automatically used as the target field.

A data source can be a dBase file that is provided with the shape file, an IBM® SPSS® Statistics data file, or an open dataset in the current session.

Figure 1. Spatial temporal prediction, select data sources
Spatial temporal prediction, select data sources
  1. For Context Data, click Add.
  2. Select Tracts_Time.sav from the list of open datasets and click Continue.
  3. In the Data and Map Association dialog, click Key Values.
  4. Select TRACTID as the data key and the map key.
    Figure 2. Spatial temporal prediction, select key field values
    Spatial temporal prediction, select key field values
  5. Click Continue.
  6. For Point Density, click Add.
  7. Select ViolentCrimeDB.sav from the list of open datasets and click Continue.
  8. In the Data and Map Association dialog, click Coordinates.
  9. For the X (longitude), select Longitude from the list.
  10. For Y (latitude), select Latitude from the list.
    Figure 3. Spatial temporal prediction, select coordinate values
    Spatial temporal prediction, select coordinate values
  11. Click Continue and then click Next.

Next