Explore and Analyze Data
Exploring and analyzing
Dimensions divide data; measures are aggregated. Discrete fields create headers, continuous fields create axes, and the blue/green pill color reflects that choice. Aggregation defaults matter: SUM, AVG, MIN, MAX, COUNT, and COUNTD answer different questions. The view's granularity is set by the dimensions on Rows, Columns, the Marks card, and active filters.
Know the filter and calculation families. Dimension filters are quick filters; measure filters are aggregate filters; context filters run early; table calculations run after aggregation and after dimension filters. Level of detail expressions come in FIXED, INCLUDE, and EXCLUDE forms. Also know groups, sets, bins, hierarchies, calculated fields, sorting, reference lines, trend lines, forecasts, and clustering.
Traps: COUNT counts rows while COUNTD counts distinct values; table calculations depend on the Compute Using setting; a measure filter cannot be combined with non-aggregated fields in the same view; discrete dates create headers while continuous dates create a time axis; Show Me changes the view, not the underlying data model.
Worked example: first order date per customer is {FIXED [Customer ID] : MIN([Order Date])}. Subtracting that from today gives tenure that is unaffected by a date filter in the view. If tenure must respect a region filter, put the region filter on Context, because FIXED ignores ordinary dimension filters but respects context filters. Rule of thumb: need one value per entity regardless of the view? Use FIXED. Need the view's granularity? Use INCLUDE or EXCLUDE.