Automating Data Validations with Board Procedures

Validating data manually before every load or report run can be time-consuming and error-prone. Board allows you to automate these checks using procedures, which can be designed to catch issues like missing values, duplicates, orphan nodes, and data mismatches before they cause problems downstream.

Using Board procedures for validation ensures consistency, speeds up workflows, and provides users with immediate feedback when something goes wrong.

How It Works

At its core, a Board procedure is a sequence of programmable steps. You can configure these steps to check entity values, cube populations, or even relationships between hierarchies and measures. These validations can run before data loads, during scheduled updates, or as part of an interactive dashboard.

For example, a validation procedure might begin by clearing a status cube, reading new data, and checking whether any codes are missing parents. If issues are found, the procedure could log them to a dedicated cube and notify users through a data view or alert panel.

Common Use Cases

One of the most frequent use cases is validating unbalanced hierarchies.

  1. Manual Validation: With a few steps, you can scan the hierarchy structure and flag records that don’t have a valid parent or are linked to a non-existent node. These are common issues that arise during financial roll-ups or department-level planning.
  2. Automated Validation: Set up a procedure that exports and imports the unbalanced hierarchy.
    1. Using the ETL function in the data reader, flag each account without a parent.
    2. Set the flag to one and load it into a cube.
    3. In another cube, check-mark the root members, which should not have a parent.
    4. You can then display these cubes in a data view and use the red alert for flagged members with a missing parent.

TIP: Turn off the Unbalanced Hierarchy for each block in this data view.

Another use case involves ensuring data completeness. A procedure might check that every expected business unit or account code is present in the latest file and raise an exception if one is missing.

You can also validate numeric inputs. For instance, if you’re loading forecast data, a procedure could flag negative values for revenue or identify unexpected zeros in expense accounts.

Best Practices

Effective validation procedures are built to be modular and reusable. A good approach is to isolate validation logic from load logic. First, run a “Check” procedure that reviews your data and populates a validation cube. Then, display the results to the user or decision-maker.

It’s also wise to make your validation results visual. You can display the outcome of your checks in a summary dashboard that shows red/yellow/green indicators for key dimensions. This makes it easy for non-technical users to understand the status and act accordingly.

Pro Tip

Use a “status cube” to store validation flags. Set its structure to align with the hierarchy you’re validating (e.g., Account or Region). This cube can be reused across multiple procedures, providing a central place to track data quality issues.

Real-World Example

A manufacturing company using Board for sales forecasting set up a nightly procedure to check for missing product mappings, duplicate SKUs, and blank sales volumes. The results are written to a validation cube, and a summary screen alerts the finance team every morning. Because of this automation, they’ve reduced reporting delays and improved trust in their forecasts.

Need Help Getting Started?

Contact QueBIT for assistance designing a custom validation workflow that fits your business model.