Each categorical column has the following number of distinct values in `raw.Website_Analytics` (62,496 rows), based on the distinct-count query already run: - **WEBSITE:** 11 distinct sites. This is the main classifier, so it is a manageable grouping for comparisons. - **YEAR:** 11 distinct values, covering 2014 to 2024. It is stored as a number but behaves like a category, one value per year. - **PAGE PATH:** 37,410 distinct values. This is very high-cardinality, close to row-level detail. - **PAGE URL:** 38,916 distinct values. This is slightly more than PAGE PATH, so some paths appear under more than one URL, for example on different sites. - **_batch_id, _source_file, _source_system:** 1 distinct value each. These are ingestion metadata, so the whole table came from a single load and carries no analytical variation. Only WEBSITE and YEAR are practical for grouping. PAGE PATH and PAGE URL are better treated as identifiers. The three metadata columns can be ignored.