Data Analysis with Surgical Accuracy and the Checklist I Use Before Any Number Goes Forward
A wrong number in a report does more damage than a missing one. People make decisions from dashboards, budgets and summaries, so every figure I hand over has to be right. Over time I have built a simple checklist that I run before any number leaves my desk.
1. Know exactly what the question is
Most reporting mistakes start before any data is touched. I write the question in one plain sentence and confirm it with the person asking. Revenue by month sounds simple until you ask whether it means invoiced, paid or booked.
2. Trace every source
I list where each piece of data comes from, who owns it and when it was last updated. If I cannot explain where a number came from, I do not use it.
3. Profile the data before analysing it
Before any formulas or queries, I look at the shape of the data. Row counts, blanks, duplicates, odd dates, text in number columns and values far outside the normal range. Ten minutes here saves hours of fixing reports later.
4. Reconcile totals with the source
Every summary should add back up to the system it came from. If my total does not match the original system, I find out why before going further. Usually it is a filter, a duplicate or a timing difference.
5. Check it from a second angle
I calculate key numbers two different ways, for example once in SQL and once in Excel or Python. When both agree, I trust the result. When they disagree, I have found a problem early.
6. Ask whether it makes sense
Numbers can be technically correct and still wrong for the business. If sales doubled overnight or headcount went negative, something is off. A quick sense check with someone who knows the business catches what formulas miss.
7. Write down the assumptions
Every report carries a short note on what was included, what was excluded and any assumptions made. It protects the reader and it protects the analyst.
Why it matters
This checklist is not about being slow. It is about being trusted. When people know your numbers are right, they stop double checking and start deciding faster. That is the real value of a good data analyst.