Solutions / BI, Report & Dashboard Testing with iceDQ

BI, Report & Dashboard Testing with iceDQ

Sandesh Gawande 10 min read

Table of Contents

Need for BI Testing

Ensure data completeness, accuracy, and consistency throughout the life cycle.

Business Intelligence Testing is defined as the process of validating the BI Report. The basis of BI Testing is reconciling the BI Reports against its metadata, validating the BI Report’s data layer /metadata against its Data Warehouse and testing the Data Warehouse against the BI source data.

The purpose of BI Testing is to present the analytical data to the Data Analyst, the Business Users, the Operational Analyst. These business users examine, and analyze this data and accordingly take appropriate and effective business decisions. These decisions form the basis for the company’s growth and success.

For the success of businesses, it becomes very important to be one hundred percent sure that these Business Intelligence Reports are accurate and reliable. BI Testing certifies your BI Report’s accuracy, uplifts the reliability of BI System and boosts the creditability of business users.

How to Automate BI Testing?

To certify the quality of the Business Intelligence Reports and automate BI Testing process, iceDQ provides reconciliation and validation object that can be customized as per your business rules. For example, the yearly sales revenue stored in the data warehouse can be reconciled with the BI report generated for it.

iceDQ is equipped to test the BI Report metadata against its equivalent database. For example, sales revenue column, year or quarter, all these dimensions and measures can be reconciled with the source data.

Now Automate Testing for Business Objects (BOXE), Tableau, MicroStrategy and IBM Cognos. With iceDQ, you can certify the reports by validating reports against business requirements or by comparing reports against database or data files.

Dashboards and reports are essential for running business. Thousands of reports are generated either daily, weekly or monthly and used for business decisions. While reports represent data, they are more like a visual component in applications. The look and feel can be visually checked but the data shown by the report is notoriously difficult to test manually. Further:

Reports visual can have graphs or tabular
Visual aspects can be changed at runtime
Reports can have runtime prompts

However, reports get their data from files or database. And reports get their data through pseudo ETL. So, if by certain means if data can be extracted from the reports then it will be theoretically tested against the database or validated against a business rule.

The Manual Approach:

Validate the report manually by looking at the report and doing some calculation in a calculator. It is difficult because the data may be rolled-up from millions of rows.

Reconcile the data by visually comparing the data in the report and then manually checking in the database. The difficulty level just multiplied by 100x now.

iceDQ’s Automated Testing of Reports:

Build a Validation Rule in iceDQ and point to an individual report or a specific report in a dashboard.
Build a Reconciliation Rule in iceDQ that points to a report or a database/file. iceDQ can then automatically compare the data and certify the report.

Thus, iceDQ not only automates report testing but also enables full automation and DevOps Implementation.

Automate the last leg of data projects with report test automation. While database provides direct access to their data, reporting tools do not. The Business Intelligence tools were primarily designed to just provide visualization. In fact, a report might be inside a dashboard and have additional filters. This makes testing reports notoriously difficult to test.