Data Certification

Data Certification is a feature that allows users to “certify” the data they have loaded, i.e. assigning a flag to data to let the admin know they are confident about them.

To enable the feature, set the environmental variable SHOW_CERTIFIED_DATA to True.

Under Admin -> SuperAdmin actions, click “Create Certified Dimension”.

This will create a filter with three items, corresponding to data not certified, partially certified, or fully certified.

The code of the filter is __certified_data, while its name, and the name of the filter items, are set by the environmental variables CERTIFIED_DATA_FILTER_NAME, CERTIFIED_DATA_FILTER_TRUE, CERTIFIED_DATA_FILTER_PART, CERTIFIED_DATA_FILTER_FALSE.

Two KPIs will also be created, with codes __CERTIFIED_DATA and __CERTIFIED_DATA_PART, type percent, and aggregation of type average.

After confirming a file, all data will be set with filter “not certified”, and the two KPIs will be created with value 0.

In the import page, two new buttons will appear under confirm options: Partially Certify Data and Fully Certify Data, corresponding to the two available levels of certification (partial and full).

When clicking Partially Certify Data, all the related data will get CERTIFIED_DATA_FILTER_PART as filter, __CERTIFIED_DATA_PART will be set to 1, and __CERTIFIED_DATA to 0 for the period/BU of the import file (this procedure may take a few seconds).

When clicking Fully Certify Data, all the related data will get CERTIFIED_DATA_FILTER_TRUE as filter, __CERTIFIED_DATA_PART will be set to 1, and __CERTIFIED_DATA to 1 for the period/BU of the import file.

After clicking either of these buttons, the button “Undo Certify” will appear.

When clicking Undo Certify, all the related data will get CERTIFIED_DATA_FILTER_FALSE as filter, __CERTIFIED_DATA_PART will be set to 0, and __CERTIFIED_DATA to 0 for the period/BU of the import file, i.e. restoring the initial status after the confirm.

After partially confirming a file, it is possible to fully certify it without undo certify, but not viceversa.

In dashboards, the filter CERTIFIED_DATA_FILTER_NAME will be available as a normal filter, so it will be possible to show only fully certified, fully and partially certified data, not certified data, and so on.

It is also possible to use the graph “Performance Single Kpi” on the two created KPIs to have a visual indication of the level of certification: for each period, the dot will be green only when all values have value 1, i.e. for __CERTIFIED_DATA when all data are fully certified, and for __CERTIFIED_DATA_PART when all data are either partially or fully certified.

WARNING: if the procedure is followed after some files have already been imported, all the related data have no filter item assigned, and the controlling KPIs have no values, so it is necessary do set all the files as not confirmed, either by hand by confirming and deconfirming them, or via a shell script calling certify_imported_values(certify=0) on all KpiImportFileHistory objects.