The Informed Company

Chapter FifteenMaintaining the Source of Truth

Now that the data warehouse is set up with an established warehouse lead, the next and ongoing step is maintenance. This involves making sure the data warehouse objects, columns, tables, views, and schemas are accurate and up‐to‐date. Maintaining a data warehouse is integral for users in an organization to easily and accurately gain insights into data. If it is not maintained, people will query the wrong data and get conflicting results.

As a company's data warehouse ages:

• New metrics need to be tracked.
• Some old metrics are no longer needed.
• Permissions will need to be granted, updated, and revoked (more than expected).
• Modeling will become un‐optimized.
• More people will be modeling.

These inevitable problems make it difficult for a company to conduct analyses. To prevent these issues, a data engineer familiar with the data warehouse becomes necessary to know how users are querying the source. This section will go in‐depth on these issues and how to address them with routine maintenance (Figure 15.1).

Figure 15.1

Track New Metrics

The ways that businesses measure their success change over time. New products launch, users behave differently, and new predictive models need to factor in new data. Success of the business depends on its ability to react to change. Sometimes this means creating a new calculated field or a new ...