Four questions before you build: the source under your report

Your department head wants a weekly report on staff occupancy from January. The data comes from your department’s scheduling tool, the same tool that kept coming back in part 2, part 3 and part 4 of Azure Basics. You build the report, IT put the source in Azure.

Before you make the first chart, you decide whether to build on this source. For that you need four answers. Each part of Azure Basics delivered one, and in this last part you lay them under a single report.

When do you call a report reliable?

When the numbers add up? When it is refreshed on time every week? All true. But… both hang on the source underneath. You ask four questions about that source: two about what is there, two about who carries it. Each question comes with a measure for a good answer, so you know when you are done.

Where does it sit, and whose space is that?

That is the tenant: your organisation’s space or a supplier’s, as part 1 explained it. Everything hangs on exactly one space, and the subscription moves along with that space. The scheduling tool runs in your own organisation’s space.

That a source works says nothing about this. The campaign site from part 1 worked too, and it sat in the agency’s space. A good answer is the name of your own organisation. Look it up for the report that gets opened most often: whose space is underneath it?

What does it consist of?

The three blocks from part 3: the place for data, the computing power and the connection. The computing power never came alone. A place for data and a connection came with it.

If someone cleans up that place because they do not recognise it as part of the tool, the whole thing stops, and your weekly report retrieves nothing anymore. A good answer is knowing which place for data came with the computing power. Would you recognise that place as part of the scheduling tool?

Who gets in?

That is what authorization from part 2 settles: the assignment of who, what and on what. You need one yourself to read the data. Who may open your report hangs on the same accounts and groups.

If that assignment is missing, you see an empty list even with a valid account. A good answer is that you and your readers each have one, on this source. Who opens your report and is about to see an empty list?

Who gets the mail when it costs more, and can that person decide?

The budget in Azure from part 4 lays the costs next to an amount once every 24 hours, mails when a threshold has been passed and stops nothing. The scheduling tool’s budget was EUR 400, and November ended at EUR 530.

The mail belongs with the owner, the person or team that decides about the tool. That owner decides whether November is worth it, and so also whether the tool keeps running. A good answer is that the mail lands with the owner and that the owner knows your report leans on the tool. Does the owner know?

Which of the four can I not answer yet?

Ask yourself that before the first chart, and put whatever stays open to whoever manages the tool. Keep the four questions for your next source after that.

In March I wrote that data governance is more than data quality, and for me, on the cloud side, it starts with these four answers. That is why cloud fundamentals are data strategy too. If your report later counts a sick colleague as available, look for the cause in the input to the scheduling tool and in the agreement on what “available” means.

Want to walk through the four questions for your own source? I’m happy to think it through with you, see trainerbjorn.nl.