Run historical reloads

Description of historical reloads of metrics.

Analytics leverages the KPI Framework in Timekeeping to calculate and load data in real time. At times it is necessary to run what is called a “historical load” for one or more metrics. This process allows the system to compute or recompute metric and KPI values for a defined historical or future period of time.

Historical loads are useful in the following situations:

• You have just licensed Analytics after using the Timekeeping application for a period of time (for example, a year) and you want to view metrics and KPIs based on that historical data.
• You have created a new metric (or UKG has introduced a new standard metric) and you want to view data for the new metric for past time periods. When a new metric has been created, the system begins computing values at the time it is saved. Values for past time periods are not computed and a historical load is required to generate data for those past periods.
• A metric has been changed and you need that change to be reflected in past or future data (for example, for scheduling and forecasting applications). Such changes may include:

  • The mapping category associated with a metric has been modified and you want the changes to apply to historical dates or future dates where the metric values have already been computed based on the original mapping category definition.

  • The data source associated with the metric has been changed. (This is a rare occurrence.)
  • The name, description, or display format of the metric has changed.

• You have loaded historical actual volume data (POS data) for use with Forecasting. The Forecasting component requires historical volume data to drive the forecasting engine. Customers often load one to three years of historical volume data when implementing the Forecasting module. This historical data can be made available through metrics and KPIs, but a historical load is needed after the creation of the metrics. This is consistent with the creation of any new metric, but it is important to note that the import of the historical volume data is a batch process that is unique to Forecasting customers. The ideal scenario would be to create the volume metrics and KPIs prior to the import of historical volume data. Using this approach, the data will be computed by the normal data flow process and no historical load will be required. If the metrics and KPIs are defined after the import of historical volume data, it will be necessary for the administrator to run a historical load onlyfor the metrics related to the volume data.

The system administrator can manually start (or schedule the start of) the historical load for one or more metrics, specifying the desired historical time period for which the load will be performed. If unspecified, the end date defaults to the current date.

In order to minimize the resources required to perform the load, the system starts the load only when those resources are not required for the ongoing real-time loads. The historical load process may be temporarily suspended. The administrator receives a message from the Control Center when the load is started and a confirmation message when the load has completed.