Explaining Master Data Quality: Evaluations, Scores, and Dimensions

Objective

After completing this lesson, you will be able to explain Master Data Quality: Evaluations, Scores, and Dimensions.

Data Quality Evaluation

Processes for Managing Evaluations

Create Process

Create an evaluation process for immediate evaluation of data quality rules.

Schedule

Schedule jobs to create and run evaluation processes. You can use this, for example, to schedule weekly evaluations.

Delete Results

Delete the results of an evaluation process to free storage space. The data quality scores of the evaluation will still be available.

Delete

Delete an evaluation process. This also deletes the evaluation results and data quality scores.

Note

Data quality evaluation uses the same process framework as consolidation and mass processing.

Data quality evaluation applies all rules to the active data in the system.

The results of the evaluation are stored for later analysis.

Schedule Evaluations

Scheduling Options
  • Date and time of start

  • Recurrence patterns (For example, daily, weekly, and so on)

Parameters
  • Description of the evaluation

  • Process template, including evaluation parameters

Creation and Configuration of Evaluations

Facts about the creation and configuration of evaluations:

General
  • Step type Evaluation (EVA) with adapter Master Data Quality Rule Evaluation is used.

  • Selection of prepared configurations (here: Evaluate with Scores) and/or using manual settings.

Evaluation Configuration

Set the Provide Scores for Dimensions indicator to include the evaluation in the score calculation for data quality dimensions and dimension categories.

Previous Evaluations
  • Delete Results – Set this indicator to remove the results of previous evaluations.

  • Delete Scores – Set this indicator to remove the scores for data quality dimensions of previous evaluations.

Customizing Implementation Guide

Transaction MDCIMGConfigure Evaluation

Note

Data quality evaluation uses the same process framework as consolidation and mass processing.

Parallelization

Prefix for Queue Name together with configuration of bgRFC allows parallel execution, load balancing, and load management.

Use and see transaction SBGRFCCONF and the respective section in the SAP Help Portal.

Evaluation Step: Results

The figure shows a result of the evaluation step. Further explanations:

Usage

Monitoring of the evaluation process

Step Messages

Contain errors and warnings during the execution of the evaluation (not errors in data)

Evaluation Results
  • Number of rule evaluations with outcome OK and with Not OK

  • Chart with distribution across the evaluated tables

The screenshot shows the Data Quality Rule Evaluation for Product in a bar chart, showing the results (OK or Not OK) or Basic Data, Description, Dimensions, and Sales Data.

Note

This is not the UI for analyzing the results. It just shows how much data is kept for this evaluation.

Scores and Dimensions

The figure here shows an example of a rule with enabled usage Data Quality Evaluation after data quality evaluation is completed.

Data quality evaluation produces several figures for each rule:

Available

The number of rows in the base table on which the rule could potentially be applied

Checked

The number of rows that were in the scope of the rule and a result is available

OK

The number of results with the outcome OK (good data)

Not OK

The number of results with the outcome Not OK (bad data)

Score

Indicator of how good data is regarding the rule: Score = OK / (OK + Not OK)

The screenshot shows the Data Quality Score on Rule level, with a score of 86.3 and status approved.

Explanations of Dimension Score Calculation:

  • Impact

    • Choose the value to reflect how important a rule is for the data quality dimension.
    • Possible values: None (0), Low (1), Medium (2), and High (3)
  • Weighting

    • The weighting of a rule's score to calculate the dimension score is calculated from the rule's impact.
    • Rules with medium impact have twice the weighting of a low-impact rule. High impact results in three times higher weighting. If the impact is none, the weighting is 0.

In the figure here of the detailed calculation for the Dimension Score, the following calculations apply:

  • Impacts: R1: 1, R2: 3, Sum: 4
  • Weighting: R1: ¼, R2: ¾
  • Dimension Score: (¼ * 39.7) + (¾ * 96.7) = 82.4 (rounded down)
The screenshot shows the screen Configure Data Quality Score Calculation for Products.

Note

The score-related cards of the Data Quality Overview app only display the category Global.

Calculation in Detail

Aggregation of rules scores to data quality dimensions:

Dimensions
  • Multiple dimensions supported
  • Score: weighted average of rule scores
  • Thresholds: Target, Warning, and Critical
Rules
  • Assigned to dimension with Impact
  • Multiple assignments possible, but only once per category
Categories
  • Multiple categories supported
  • Special: Global category used for data quality overview
  • Score: Average of dimension scores
  • Thresholds: Target, Warning, and Critical
The screenshot shows the Data Quality Screen with status completeness and a score of 82.4. Target is 85.

Special Cases with Scores of Zero

The following explains the dimensions used in the figure, which shows special cases with scores of zero:

Impact None → Weighting Zero

Rule assignment R1:

Impact is None → Weighting 0 → Does not influence dimension score

Rule Score is (really) 0

Rule assignment R2:

No product complies to this rule.

Rule Score is undefined
  • No outcomes of rule evaluation available for the following reasons:
    • No data in scope
    • Rule not yet evaluated
  • Rule assignment R5: handled as if impact is None

In the preceding figure showing special cases with scores of zero, the following calculations apply:

  • Impacts R1: 0 / R2: 2 / R3: 1 / R4: 3 / R5: 0 Sum: 6
  • Weighting R1: 0 / R2: 2/6 R3: 1/6 / R4: 3/6 / R5: 0
  • Dimension Score: (0* 100) + (33.3% * 0) + (16.6% * 39.7) + (50.0% * 96,7) + (0%*0) = 54.9
The screenshot shows the Data Quality Dimensions screen with the special case score of zero.

Data Quality Analysis

The figure of the Data Quality Evaluation Overview and the available cards illustrates the following:

Latest Data Quality Score
  • Data quality score by dimensions of the category, Global
  • Drill-down into the results to analyze the score by dimension, rule, and so on, down to the actual rule outcomes
Data Quality Trend
  • Trend of the dimension scores of the category, Global
  • Drill-down
Incorrect Data (3 cards)
  • Various views (for examples, by product status, by plant)
  • Drill-down
The screenshot shows some available cards for Data Quality Evaluation.

The figure of the evaluation results illustrates the analytical list page.

All apps implement the SAP Fiori analytical list page (ALP) floorplan and share the same behavior. Refer to https://experience.sap.com/fiori-design-web/analytical-list-page/.

The following numbered items correspond to the UI shown in the preceding figure:

  1. Header with the filter bar, in the preceding figure, displaying visual filters for Results by Plant and Results by Plant-Specific Product Status

  2. Evaluation Items chart - smart chart for visualization, drill-down, and filtering

  3. Evaluation Items table - analytical table for detailed information, actions, and export

The screenshot shows the Evaluation Results for Plant Data of Products screen.

Whatever you define in the filter bar influences the data displayed in the chart and in the table.

You can also switch between visual filter and compact filter.

You can add more filters with the Adapt Filters button.

The screenshots shows the filter bar and the influence of the displayed data.

Note

The Evaluation Results for … tiles navigate to the apps with the filter for the outcome that is set to Not OK.

The smart chart allows you to define your own chart type, choose dimensions from the product master data, and drill down to see the result by the values of almost any field of the product basic data section.

The screenshot shows the Evaluation Items in a bar chart.

The table allows you to add fields from the product basic data section as columns. This and further functions are available in the settings dialog.

The screenshot shows the table view and the settings to add or remove colums to the table.

Data Quality Score for Products

Usage

  • You want to see, compare, and analyze the scores of all data quality dimensions (and not only those of the dimension category, Global).
  • You can further drill down into the scores from here, down to the actual evaluation results by rule.
The screenshot shows the Dimension Category bar chart in the Data Quality Score For Products UI.

The figure shows an example of a view on product-level by item category group.

The screenshot shows an example view of the Evaluation Results for Products. The UI consists of 3 parts. The Filter area on top, the Evaluation Items as bar chart in the middle and in the lower part the table view.

Data Quality Remediation

To correct erroneous data, perform the following steps:

  1. Delegate work (share page).

    Send the link to the page including filters to delegate correction work.

  2. Export of evaluation items.

    Export the table with the evaluation results in OpenOffice (XLSX) format.

  3. Navigate to all apps in the role of the user for single products or business partner.

    Examples: open fact sheet, change with change request, and so on

  4. Process selected objects with mass processing for multiple products or business partner.

    Select at least one row.

  5. Export products or business partners.

    Select at least one row.

    Export selected objects in OpenOffice (XLSX) format for offline editing and later import in the Mass Processing app.

The screenshot shows the option to correct erroneous data.

Data Provider Integration

The data provider integration using the CDQ platform and its services requires central governance in cloud-ready mode.

The figure explains the data provider integration to increase data quality.

Fundamentals

  • Data provider integration is an extension of data quality management capabilities of SAP Master Data Governance.
  • SAP Master Data Governance acts as a gateway to access external master data content and services provided by data provider.
  • Customer value is added by integrated use cases, leveraging external services, optimizing and automating end-to-end master data management processes.

Commercial Aspects

  • The usage of data provider integration is included in SAP Master Data Governance.
  • An additional contract with the data provider is required.
The figures explains the integrations using HTTPS and REST between SAP Master Data Governance on S/4HANA and CDQ Cloud Platform.

Features

  • Users can look up reference data to create and enrich business partners.
  • CDQ provides access to numerous data sources.
  • The availability of the CDQ platform, its services, and data sources is subject to your agreement with CDQ AG and independent of contracts with SAP SE.
  • A description of the CDQ Business Partner LookUp service and its configuration is provided by CDQ.
  • It requires a contract with CDQ AG: See SAP Store.

Evaluate the Master Data Quality of Material

Business Example

You are a master data quality specialist. Your task is to maintain master data quality rules and evaluate the data quality based on the rules for products.