**The categorical columns are either constant or nearly uniform, so no value dominates.** | Column | Distinct values | Most common values | |---|---|---| | `company_name` | 152 | ASM International, BE Semiconductor and Soitec (3,475 rows each, 0.82%) | | `stock_name` | 152 | ASM.AS, BESI.AS and SOI.PA (3,475 rows each, 0.82%) | | `_batch_id` | 1 | `batch_d9e1fae7c9b5` (100%) | | `_source_system` | 1 | `api` (100%) | | `_source_file` | 1 | the single Kaggle `semi_conductor_se.csv` (100%) | - **Company and ticker:** `company_name` and `stock_name` map one-to-one. Each ticker pairs with one company in the top-value results (for example, Soitec is SOI.PA and Infineon is IFX.DE), and both columns have 152 distinct values. - **Next most common:** Melexis NV (MELE.BR) has 3,473 rows. Elmos Semiconductor, Infineon and SÜSS MicroTec (each 3,448 rows) come next, then Aixtron and centrotherm international (3,445 rows each). All of these sit at about 0.82% or 0.81% of the 422,729 rows. - **No skew:** The most frequent company has only slightly more rows than the 15th-ranked one (3,475 vs. 3,415), so no company dominates. The grain is one row per stock per trading day, so row counts mostly reflect how long each ticker has traded. The leaders all start on 2012-01-02 and end on 2025-08-01. The 15th-ranked, Analog Devices (ADI), starts on 2012-01-03. - **Different end date:** Weebit Nano and BrainChip (both listed in Australia) end on 2025-08-04, slightly later than the others. - **Metadata columns:** The three `_batch_id`, `_source_system` and `_source_file` columns each hold a single value, so they carry no analytical information. The earlier dataset card said roughly 149 distinct companies, but the actual count from the query is 152. All figures come from the frequency and date-range queries I ran on `raw.semi_conductor_se`.