**The strongest relationship in this dataset is between a company's age (`Founded`) and its funding stage (`Stage`).** Funding amount (`Amount`) is also moderately tied to stage. Most other columns are free text or identifiers, so they can't be correlated this way. **Correlations (step-2, 426 rows, with 385 usable amounts and 256 usable stages)** | Pair | Pearson r | Reading | |---|---|---| | `Founded` vs `Stage` | **-0.54** | Strongest. Older companies are at later stages. | | `Stage` vs ln(`Amount`) | **+0.40** | Later stages raise bigger rounds. | | `Founded` vs ln(`Amount`) | **-0.25** | Weak. Older companies raise somewhat more, mostly because they are at later stages. | The `Amount` correlations use the log of the amount, because amounts are very skewed. **Stage breakdown (step-1)** - Median round size rises steadily with stage. It is $0.5M for Pre-seed, $1.6M for Seed, $12M for Series A, $22M for Series B, $60M for Series C and $71M for Series D. It then jumps to about $316.5M for Series F and $364M for Series E. - Average founding year falls with stage, from about 2021 for Pre-seed and 2019 for Seed to about 2012–2014 for Series D–F. This matches the -0.54 correlation. - The "undisclosed" stage group has a median amount of $23.5M and an average founding year of 2003. These are probably older, established companies, which weakens the overall correlations. **Caveats** - **Messy stage labels (step-3):** `Stage` has variants and misplaced values, such as "Preseed" vs "Pre-seed", "Undiclosed", "$Series A", and investor names or amounts sitting in the stage column. Only 256 of 426 rows had a usable stage, so the stage correlations are approximate. - **Text amounts (step-0):** `Amount` is stored as text. It includes "Undisclosed" (35 rows) and blanks (25 rows), plus some malformed values like "$165.5". Only 385 rows had a usable amount. - **May 2022 table:** It has misaligned rows, and its `Founded` column is text rather than a number. - **Observational data:** These are associations only. The data cannot show that stage causes larger rounds or the reverse. - **Not tested:** I did not measure how `Sector` or `Location` relate to the numeric columns, so I can't say whether they are linked to amount or stage.