Derived columns carry the analysis forward: ratio, difference, category. Compute them once, store them, and reuse. Recomputing the same expression in five places is how five subtly different answers end up in one report.
About 2 min of reading
Add a share column and check it sums to one.
Compute the share
Before you run it, say out loud what it will print. Then open the answer.
Prove it is consistent
Before you run it, say out loud what it will print. Then open the answer.
Before you write anything, put these lines back in the right order. Click a line to move it up one step. This support step is easier than writing code from scratch, and it is here to make the last example stick.
Read the first example one line at a time. A line that prints shows what appears on screen; the rest set things up quietly.
A column called 'share' beats 'x2'. When the column name states the question it answers, the analysis reads like an argument and reviewers can check each step against the claim it supports.
If two places compute 'is_over_180_days', one of them will eventually drift. Put the rule in one function and call it. The bug you are looking for is not a wrong number, it is two different numbers.
for r in rows: r['share'] = int(r['count']) / total. Now every later step sees the same value. In JavaScript, rows.forEach(r => r.share = r.count / total) or map to new objects.
Division by zero from an empty group is a crash or an infinity that prints as 'inf' and quietly makes your total wrong. Test the denominator and decide what the answer should be when it is zero: usually 0 or None, never a silent infinity.
Rounding a derived column early means the parts no longer add up to the whole. Keep full precision through the pipeline and round only when you print. This is also why percentages should be reported as fractions of the same denominator.
Each one is defined here on this site. Open the entry for the full definition and the other words it points at.