**Only two columns are numeric:** `Amount in USD` (stored as text, but 2,065 of 3,044 values parse as numbers) and `Sr No` (a row ID). The other columns are text or timestamps. The figures below come from the profiling query in step 1. **Amount in USD (2,065 numeric values of 3,044 rows)** - **Range:** the minimum is $16,000 and the maximum is $3.9 billion. - **Mean:** about $18.4 million. - **Median:** $1.7 million. - **Quartiles:** the 25th percentile is $470,000 and the 75th is $8.0 million. - **95th percentile:** $60 million. - **Spread:** the population standard deviation is about $121 million, roughly 6.6 times the mean. The distribution is extremely right-skewed. The mean is about 11 times the median, and the interquartile range ($0.47M to $8M) is tiny next to the maximum. A few mega-deals pull the mean up, so the median and percentiles describe a typical deal better than the mean does. About 979 rows (32%) have no parseable amount, so these statistics cover only the two-thirds of deals that report one. Two of my attempts to compute the standard deviation errored with "out of range", most likely because of extreme or malformed values. The standard deviation shown comes from the version of the query that succeeded (step 1), which is labelled `stddev_pop_lt_1e12`, so I take it to be restricted to values under 1e12. I did not run a separate check to confirm that. **Sr No (3,044 values, none missing)** - **Range:** 1 to 3,044. - **Mean and median:** both 1,522.5. - **Quartiles:** 761.75 and 2,283.25. - **Standard deviation:** about 878.7. This is a sequential identifier, a uniform run from 1 to 3,044. Its distribution carries no analytical meaning.