Method

A DAX measure is a claim about a number. This site treats it that way.

What every guide carries

• The dataset. A CSV you can download, with the row count and the checksum stated on the guide page.
• The measures in full. A downloadable measures file with the exact DAX, so there is nothing to retype from a screenshot.
• The expected results. A table of the values the measures must return, month by month, so you can tell in one glance whether your report agrees.
• Sources. Microsoft Learn pages for every DAX function and every Desktop step the guide relies on.

How the numbers are verified

Each measure's result is computed a second way: the identical aggregation is run in Python (pandas) against the same CSV file, and the guide publishes only when the two agree exactly. The Python version and the measured results are printed on the guide page. The charts on the guide pages are drawn from those same computed numbers.

What was not run

Power BI Desktop runs on Windows only, and it is not installed on the machine these guides are built on. The Desktop interface steps — importing the CSV, building the relationship, adding the visuals — are documented from Microsoft's own docs and cited, but they were not clicked through here. Each guide says this on the page. If you follow a guide and a Desktop step behaves differently from what is written, that is a real discrepancy worth reporting, not a detail to work around silently.

About the data

The sales data in the launch guides is a constructed sample: three product lines, monthly revenue from January 2024 to October 2025, 66 rows. It is not a real company's sales. It exists so the measures have something concrete to compute against, and so every number on the page can be re-derived from the file you download.