Nearly every column has some missing values, but none of them are true SQL NULLs. The checks in step-0 and step-12 found zero NULLs in company, stage and amount. The gaps are blank or empty-string cells (stage appears as `[]` in step-3). The only columns that are fully populated are none, so the lowest rates are about 4%. **Overall missing rate across all 446 rows** (counts from the "ALL" row in steps 9 and 10; the percentages are those counts divided by 446): - **Stage:** 127 missing (28.5%), by far the worst column. - **Investor:** 45 missing (10.1%). - **Founders:** 38 missing (8.5%). - **Amount:** 25 missing (5.6%). - **Location:** 20 missing (4.5%). - **Company Name and Sector:** 19 missing each (4.3%). - **Founded and What it does:** 18 missing each (4.0%). **Where the gaps are concentrated** (per-month rates in step-4): - **Stage** is mostly a January–March problem. It is blank for 37.4% of January rows, 30.2% of February rows and 31.6% of March rows, but only 3.2% of April rows. May is 50.0%. - **May** is the worst month, with about 43–64% missing in every column. For example, 45.2% of company names, 50% of amounts and 64.3% of founders are blank. The dataset card notes that some May rows are misaligned or blank, so these look like structural blanks (about 18–19 rows empty across most columns) rather than genuinely unreported values. - **January to April** are otherwise nearly complete in company, founded, location, sector and "what it does". The only exception is a single blank location in April (1.1%). - **Founders** is missing for 10.5% of April rows (10 rows) and 1.0% of February rows (1 row), and none in January or March. - **Investor** is missing for 3–10% of rows in each of January to April. - **Amount** is missing for 2 rows each in February and April (about 2%), and none in January or March. Excluding May, the overall problem columns are stage, investor and founders. Any stage-based analysis would lose about 28% of rows, or roughly a third of the January–March data.