Only two columns across the five monthly tables are numeric in any useful sense: **Founded** (a year) and **Amount** (stored as text like "$1,000,000", so I parsed it to USD). The figures below pool all 446 rows from the January–May 2022 tables (step-1 query). | Metric | Founded (year) | Amount (USD) | |---|---|---| | Valid values | 427 of 446 | 385 of 446 | | Min | 1871 | 11.3 | | 25th percentile | 2014 | 1,300,000 | | Median | 2017 | 9,400,000 | | Mean | 2015.3 | about 145.6 million | | 75th percentile | 2020 | 45,000,000 | | Max | 2022 | 18.3 billion | | Std deviation | 11.0 | about 991.5 million | **Founded** - The distribution is left-skewed. The mean (2015.3) sits below the median (2017) because a few very old companies pull it down, with the oldest founded in 1871. - The middle half of companies were founded between 2014 and 2020, so most are young startups. - The standard deviation of 11 years is inflated by those old outliers. The interquartile range is only 6 years. - 19 rows have no usable year. May's `Founded` column is text and contains misaligned rows. **Amount** - The distribution is extremely right-skewed. The mean (~$145.6M) is about 15 times the median ($9.4M), and the standard deviation (~$991M) is roughly 7 times the mean. A few mega-rounds dominate the average. - The middle half of rounds falls between $1.3M and $45M. The range runs from about $11 to $18.3 billion. - The $11.3 minimum and $18.3 billion maximum look like data-entry or unit problems rather than real rounds. Check them before relying on the extremes. - 61 rows have no numeric amount. Many of these are "Undisclosed", which appeared 6 times in the value counts (step-0). The rest are blank or misaligned rows, mainly in May. - For typical funding size, use the median and IQR rather than the mean.