### 1.1 Business Data Landscape This dataset is a professional men's tennis performance archive, sourced from Kaggle. Its 5 tables hold 718,470 rows. Three large tables (raw_kaggle, return_kaggle, serve_kaggle) each hold roughly 237,000 rows. Their near-identical counts suggest they describe the same underlying set of matches from different angles: raw match records, return performance, and serve performance. Two smaller reference tables, players_man_ (462 players) and players_tournament_man_ (6,422 rows), supply the player and tournament context around those records. ### 1.2 Key Data Assets | Table | Role | Rows | % of Total | Size | |-------|------|-----:|----------:|-----:| | raw_kaggle | path | 237,205 | 33.0% | 0.1 MB | | return_kaggle | path | 237,196 | 33.0% | 0.1 MB | | serve_kaggle | path | 237,185 | 33.0% | 0.1 MB | | players_tournament_man_ | dimension | 6,422 | 0.9% | 0.0 MB | | players_man_ | dimension | 462 | 0.1% | 0.0 MB | The three path tables together account for 99% of all rows. They form the performance core of the dataset: match-level detail split into raw, serve, and return views. The two dimension tables are compact by design, serving as the lookup layer for players and player-tournament participation. The row counts of the three performance tables differ by only 20 rows (237,185 to 237,205). That is a useful signal of consistent coverage, and it may be worth understanding what explains the small gap. ### 1.3 Data Integration Posture Nettle detected 0 validated joins, so the five tables currently stand as separate assets rather than a connected model. Referential integrity scores 100%, which means there are no broken links to repair, but it also reflects that no links were established to test. For analytics, this means questions that span serve, return, and player context, such as comparing a player's serving strength with their return game, cannot yet be answered directly from the structure as delivered. Such questions would first require confirming which shared identifiers, if any, the tables hold in common.