Data Exploration and Analysis
Fields, aggregation, and level of detail
You must be fluent in dimensions vs measures and the discrete (blue) vs continuous (green) distinction - the same field can be either, and that choice changes whether you get headers or axes. Know the aggregation functions: SUM, AVG, MIN, MAX, COUNT, COUNTD, and ATTR (returns * when multiple values exist). When a dimension is in the view, every measure is computed at that level of detail; two-pass aggregation (SUM(SUM([Sales]))) lets you aggregate an aggregate.
Calculated logic
Level of Detail expressions - FIXED, INCLUDE, EXCLUDE - set their own grain. FIXED is computed against the data source before dimension filters, so a FIXED total ignores ordinary filters unless they are context filters. Table calculations (running total, percent of total, rank, WINDOW_AVG) run on the aggregated results in the view and depend on Compute Using (addressing and partitioning).
Traps to watch
- Filter order: extract → data source → context → dimension → measure → table calculation. A table-calc filter removes marks only after aggregation.
COUNTDvsCOUNT: duplicates matter.- Quick filters vs parameters: parameters are constant until changed and cannot be filtered on directly.
- Hiding a field is not the same as excluding it from the data source.
Worked example
You need each region's share of the global total, unaffected by a country filter. Use { FIXED : SUM([Sales]) } as the denominator: SUM([Sales]) / { FIXED : SUM([Sales]) }. If the filter must still apply, first add the filter to Context. By contrast, Quick Table Calculation → Percent of Total with Compute Using set to Table answers a different question - share of whatever is currently in the view.