Scoring definitions - Data360_Govern - Latest

Data360 Govern Help

Product type
Software
Portfolio
Verify
Product family
Data360
Product
Data360 Govern
Precisely Data Integrity Suite > Govern
Version
Latest
Language
English
Product name
Data360 Govern
Title
Data360 Govern Help
Copyright
2024
First publish date
2014

A score is a percentage that represents the health of one aspect of an asset, based on a set of measures that you define.

Measures can be organized into groupings, and can include conditions, which determine whether or not the measure is applied to a particular asset. Scores are calculated for each measure that you define. The scores from these measures are then combined to calculate the overall score for an asset. For example, you might create a governance score type with one set of measures for Business Terms, and then create a second governance score type with a slightly different set of measures to assess the health of Applications.

Note: The Scoring tab and scoring badges are only visible on assets where a scoring definition has been defined and calculated.

Creating scoring definitions

  1. Navigate to Configuration > Scoring Definitions.

  2. Click the Create Scoring Definition button to create a new score.

  3. In the Create Scoring Definition dialog, select a Score Type. Choose from:

    • Governance
    • Data Quality
  4. Select the Asset Type with which you want to associate the score. You can type to filter the list of asset types.

  5. Select Internally calculated or Externally calculated. The default is Internally calculated for both Governance and Data Quality score types.

    If your score type is internally calculated, the overall score is calculated within Data360 Govern, taking into account any conditions that are set on the measures. If it is externally calculated, all scores and measures must be passed in by an external provider and you cannot set any conditions on the measures. For example, you can use Data360 DQ+ to calculate a score and then pass the calculated score to Data360 Govern through the API. For more information, see Internally and externally calculated scores.

  6. Click and drag the sliders to define Good (green), Average (amber) and Poor (red) scoring bands. You can define any percentage range for each band. The color of a score badge on an asset reflects the scoring band within which the score value sits.

  7. Click Create.

Adding measures

  1. In Configuration > Scoring Definitions, click the row that contains the scoring definition to which you want to add the measure.

    Note:

    If the score type is internally calculated, you must add at least one measure to complete the scoring definition. However, internally calculated data quality measures require relationships to exist between the asset type being scored, and rule types.

    If this is not the case, a message will be displayed, stating that before you can create any measures, you need to define at least one relationship type to associate the asset type with rule types, either directly or through relationships to other asset types. Until this relationship is established, the Add Measure button will be disabled. See Defining relationships between assets and Relating quality rules to assets for more details.

    If the score type is externally calculated, you can choose whether or not to add measures. If you do define measures, the only fields that are available for externally calculated scores are Name, Description and Effective Date. For more information, see Internally and externally calculated scores.

  2. To add a new top-level measure, click the Add Measure button at the top right of the Measures panel.

    Or, to add a child measure, click the Add measure to group button to the right of an existing parent measure. The Add measure to group button is only available on measures that are set as a Grouping Measure.

    The Create Measure dialog opens.

  3. Enter a Name for the measure.

  4. Enter a Description for the measure.

  5. Choose an Effective Date for the measure using the calendar widget, or enter a date in the format mm/dd/yyyy, for example 02/10/2020. This is the date from when the measure will be used in score calculations.

    Future effective dates are allowed. When used, and depending upon the type of measure that has been created, this is the date from which the score will be calculated.

  6. Enter a percent value for the Weight of the measure, between 1 and 100.

    The weight of a measure expresses the importance of the measure relative to another. Weighting is applied to groupings, and to measures contained within groupings. For example, a measure with a weight of 50% will contribute twice as much as a measure with a weight of 25%. The sum of all measures used to calculate the score should sum to 100% (at the top-level with and within each group of measures). Where they do not, they will be adjusted when used to calculate the score.

  7. If you are creating or editing a measure for an internally calculated data quality score, you have the option to use a Pass Threshold, or not. If this option is available, a tooltip displays information about how the threshold works. Do the following:

    1. Select Yes if you want to use a pass threshold.

      The Pass Threshhold (%) field displays.

    2. Enter a threshold greater than zero and less than or equal to 100.

      Decimal numbers can be entered, but only to 3 decimal places and, if you make a mistake, an error message displays. If the pass threshold is used, a pass threshold override is added to condition groups.

      You can see if a threshold has been set, together with its value and impact in the Calculation and Definition tabs of the Scoring side panel.

  8. You can use groupings to collect related measures together and calculate an overall score for the group. If you want to create a "Grouping Measure" that will contain other measures, select Yes on the Grouping Measure property. In this case, the score is calculated based on the child measures within the group. For more information, see Groupings.

    Note: Only top-level measures can be defined as a group. It is not possible to define a child measure as a group.

The next steps depend on whether you are adding a measure for a governance score, or a data quality one. See Internally calculated governance measures, or Internally calculated data quality measures, as appropriate, for further details.

Internally calculated governance measures

  1. For internally calculated governance scores, select a Test Type to determine the type of check to perform to establish whether a measure has passed or failed. Choose from:

    • Field - Checks for the presence of a field, or if a specified value exists for a field. Depending on the selected field type, various operators can be used to define the test criteria. The following field types are supported:
      Field type Operator Description
      Date

      is

      The measure will pass if the date field value matches the specified date.

      is not

      The measure will pass if the date field value does not match the specified date.

      is before

      The measure will pass if the date field value is before the specified date.

      is after

      The measure will pass if the date field value is after the specified date.

      is populated

      The measure will pass if the date field contains a value.

      is not populated

      The measure will pass if the date field does not contain a value.
      Date With Time

      is populated

      The measure will pass if the date with time field contains a value.

      is not populated

      The measure will pass if the date with time field does not contain a value.
      List

      is

      The measure will pass if the list field matches the selected value. You can only choose one value.

      is not

      The measure will pass if the list field does not match the selected value. You can only choose one value.

      is populated

      The measure will pass if the list field contains a value.

      is not populated

      The measure will pass if the list field does not contain a value.
      Number or Decimal Number

      is

      The measure will pass if the number field matches the specified value.

      is not

      The measure will pass if the number field does not match the specified value.

      is greater than

      The measure will pass if the number or decimal number field has a value greater than the specified value.

      is greater than or equal to

      The measure will pass if the number or decimal number field has a value greater than or equal to the specified value.

      is less than

      The measure will pass if the number or decimal number field has a value that is less than the specified value.

      is less than or equal to

      The measure will pass if the number or decimal number field has a value that is less than or equal to the specified value.

      is populated

      The measure will pass if the number or decimal number field contains a value.

      is not populated

      The measure will pass if the number or decimal number field does not contain a value.

      Simple Text

      is

      The measure will pass if the simple text field matches the specified value. The specified value is not case sensitive and leading or trailing spaces will be trimmed, but not spaces in between words.

      is not

      The measure will pass if the simple text field does not match the specified value. The specified value is not case sensitive and leading or trailing spaces will be trimmed, but not spaces in between words.

      contains

      The measure will pass if the simple text field contains the specified value. The specified value is not case sensitive.

      does not contain

      The measure will pass if the simple text field does not contain the specified value. The specified value is not case sensitive.

      starts with

      The measure will pass if the specified value matches the beginning of the simple text field.

      ends with

      The measure will pass if the specified value matches the end of the simple text field.

      is populated

      The measure will pass if the simple text field contains a value.

      is not populated

      The measure will pass if the simple text field does not contain a value.
      Html/Richtext

      is populated

      The measure will pass if the Html/Richtext field contains any text or images.

      is not populated

      The measure will pass if the Html/Richtext field does not contain any text or images.

      True/False

      is true

      The measure will pass if the field is true.

      is false

      The measure will pass if the field is false.

      is populated

      The measure will pass if the true/false field contains a value.

      is not populated

      The measure will pass if the true/false field does not contain a value.
      Note: If you want to use a value such as 10 in a decimal field, you should use operators that allow a range of 9.99 to 10.01 to be specified. For example 'greater than' and 'less than'.
      Note: If a custom field that is used in a pass test is deleted from the asset that is being scored, the measure is ignored.
    • Relationship - Checks if a relationship of the selected type exists on the asset. You can only choose from relationships in which the asset type of the score is the subject or object. The measure will pass if at least one relationship of the selected type exists on the asset.

    • Predicate - Checks if a relationship based on the selected predicate exists on the asset. The predicate test provides a broader check than the relationship test, as it checks for all relationship types that use the specified predicate, rather than checking for a specific relationship type. The measure will pass if at least one relationship exists on the asset where the relationship type uses the selected predicate.

      Both Relationship and Predicate checks can be used to determine whether a parent has a child asset type or not, but not the reverse. They can be applied to both business and technical assets.

      You can also define Test Type Relation or Test Type Predicate measures for both parent to child relationships and predicates, and apply them to models and policies as well, if needed.

      The following example focuses on a business or technical asset. Do the following:

      1. Select Relationship or Predicate as the Test Type.
      2. Select the appropriate Test Condition. For example, to check the presence or not of any children of the MB Table asset type, you might select Parent Of MB Table/MB Column.
      3. Select either the exists or does not exist operator.
    • Ownership - Checks if an owner based on the selected responsibility type exists on the asset. The measure will pass if at least one owner is assigned the selected responsibility type. The assigned owner can be an individual, a group or an organization.
      Note: If a group or organization is assigned, the measure will pass regardless of whether there are users in the group or organization.
      Note: If a responsibility type is removed from the asset type that is being scored, the measure is ignored.
    • External - If your measure cannot be tested by selecting one of the other options, select External to use an external system to calculate the score, such as Data360 Analyze or Data360 DQ+. Optionally, you can enter a value in the Instruction String field to automate scoring with Data360 Analyze or Data360 DQ+.
      Note: External should only be selected where advanced scoring checks are required. In most cases, one of the other test type options should be used.

      Note that the Test Type field is not displayed on top-level grouping measures, as in this case, the score is calculated based on the child measures within the group.

  2. If required, you can add conditions and their appropriate weighting to a measure, to determine if and how the measure is applied to an asset. For more information, see Asset conditions and weighting.

  3. Click Create.

Internally calculated data quality measures

  1. For internally calculated data quality scores, make a Rule Results Selection.

    If there is only one option available, it is selected by default. Otherwise, a scrollable list is displayed.

    Note: The list may be made up of complex rule results, such as, for example, Application calls Procedure, which executes Function, which returns Function Column, which is evaluated by Rule.
  2. Select the appropriate Results Operation:

    • Average (default).
    • Maximum.
    • Minimum.
  3. Select Rule Result Filters, if required.

    Rule result filters are applied to fields supplied by rules and any intermediate asset types that are used to relate rules to your scoring asset type.

    Fields are listed for any asset types featured in the Rule Results Selection relationship chain, excluding the asset type being scored. Multiple asset types such as Field, Table and Rule are also available to filter on, and can be quickly identified by entering an appropriate search string.

  4. Select the required operator.

    For example, less than or equal to, is not, or similar. The list of operators and the default choice varies according to the selected filter.

  5. Enter a value.

  6. Add further filters as required.

  7. If required, you can add conditions and their appropriate weighting to a measure, to determine if and how the measure is applied to an asset. For more information, see Asset conditions and weighting.

  8. Click Create.

    The Create measure dialog closes and the measure is added to the score definition page.

The data quality score is calculated whenever rule results are loaded.

For score calculation examples, see Internally calculated data quality score examples.

Editing or deleting scoring definitions

  1. Navigate to Configuration > Scoring Definitions.

  2. To edit a scoring definition, click the Edit button to the right of the score that you want to edit.

    Alternatively, click the row that contains the scoring definition that you want to edit. This will open a page showing the Score Definition and Measures. From here, you can edit the scoring definition, or add and edit measures.

    To delete a scoring definition, click the Delete button to the right of the score that you want to delete.

    Note: You can only edit and delete scores for which you have not defined measures.

Editing measures

  1. Navigate to Configuration > Scoring Definitions.

  2. Click the row that contains the scoring definition for which you want to edit a measure.

  3. Click the menu button to the right of the measure that you want to edit and select Edit.

    The Edit Measure dialog opens.

  4. Edit the measure properties as required.

    Note: Any changes you make to the measure will result in a new version of the measure being created. You must provide a new Effective Date, unless you are changing the weight of the latest effective date for a measure and there are no results associated with that date. The effective date is the date from when the changes should take effect. For example, if you need to adjust the weight value for a measure with existing calculations, the new weight is applied to any score calculations from that date forward, but scores that have already been calculated will not be affected.

    Future effective dates are allowed. This will be the new date from which the score will be calculated.

  5. Click Save Changes.

    This button is disabled until a change is made.

  6. To close the dialog without making changes, click Close.

    This button changes to Discard Changes when a change is made.

Disabling a measure

  1. Navigate to Configuration > Scoring Definitions.

  2. Click the scoring definition which contains the measure that you want to disable.

  3. Click the menu button in the corresponding row in the Measures panel and select Disable.

  4. In the Disable Measure dialog, click Disable.

Note: When you disable a measure, it is no longer used in score calculations, but any results that have previously been calculated using the measure are retained in the version history for the measure. If you disable a grouping measure, all measures grouped within that measure will also be disabled.

Viewing disabled measures

When a measure is disabled, it is no longer visible by default. To display disabled measures:

  1. Navigate to Configuration > Scoring Definitions.

  2. Click the scoring definition that contains the relevant measures.

  3. Select the Show disabled measures check box.The disabled measures display, along with a Disabled badge to the right of the name.

Viewing measure version history

When a measure is edited and it has been used in a score a new version is created. To view the version history:

  1. Navigate to Configuration > Scoring Definitions.

  2. Click the scoring definition which contains the measure that you want to disable.

  3. Click the menu button in the corresponding row in the Measures panel and select Version History.

    The Version History dialog opens, listing each version on the left and the properties for the selected version on the right.

    You can select a previous version to see what has changed between versions.

  4. Click Close to return to the Scoring page.

Groupings

You can use groupings to collect related measures together and calculate an overall score for the group.

For example, you have a set of advanced measures and a set of basic measures that you want to group together as follows:

Measure grouping

You can apply different weightings to the groups to determine their relative importance, and similarly you can apply weightings to the individual measures within each group. For more information, see Weighting.

To group measures:

  1. Create a new top-level measure by clicking the Add Measure button at the top right of the Measures panel. Or, to edit an existing top-level measure, click the menu button to the right of the relevant item in the table and select Edit.

  2. Select Yes on the Grouping Measure property.

    A measure that is defined as a group will have an Add button in its row of the Measures table. This indicates that you can add child measures to the group.

    Note: Only top-level measures can be defined as a group. It is not possible to define a child measure as a group.
  3. Add child measures to the group by clicking the Add button in the table, to the right of the top-level measure.

You can add the same measure to different groupings. Measures that are part of more than one grouping are independent of each other, and are scored as such. This means that you can have two measures in different groupings with the same name, and they will be treated completely separately, so you can apply different weightings and different conditions to each measure.

Note: Once you have created a measure as a grouping you cannot change this value, even if you delete all the measures in the grouping.

Internally and externally calculated scores

When a score is calculated internally, the results of the measures are passed into Data360 Govern by an external provider, but the overall score is calculated within the application, taking into account any conditions that are set on the measures. For an externally calculated score, the results of any defined measures, and the score itself, are passed into Data360 Govern by an external provider.

It is not mandatory to define measures on an externally calculated score, but if you do, the only fields that are available are Name, Description and Effective From. You can select Externally Calculated on an existing score type if there are no measures defined. You cannot set any conditions on a measure of an externally calculated score.