Each monthly table has its own distinct counts. I did not compute a combined count across all five months, so the totals below are per-table only. These come from the distinct-count query in step 0 (the same result came back in step 1). **Company Name** is almost unique within each month. January has 114 distinct names in 115 rows, April has 94 in 95, and February and March are fully unique (96 and 98). May is the outlier, with only 24 distinct names in 42 rows. The dataset card notes that some May rows are misaligned or blank. Since `COUNT(DISTINCT)` skips nulls, the blank rows probably explain the gap. **Location** is low-cardinality, with 11 to 17 distinct values in January through March. April jumps to 39, and May has 16 (in a much smaller table). The April figure may reflect messier or more granular location entries. **Sector** is the most varied classifier. It has 42 distinct values in January, 39 in February, 54 in March, 58 in April and 19 in May. With 40 to 58 sectors across roughly 100 rows, sector labels are fragmented and would need grouping before any analysis. **Stage** is compact, with 12 distinct values in January, 11 in February, 11 in March, 18 in April and 6 in May. **Investor** is the highest-cardinality column after Company Name, with 107 distinct values in January, 86 in February, 90 in March, 54 in April and 22 in May. Many rows probably list multiple investors in one string, so these counts are of distinct strings, not distinct investors. April stands out on Location and Stage, and May is distorted by blank or misaligned rows. Treat distinct counts across months as only roughly comparable.