Analyze Data to Answer Questions
Core concepts
Analysis means aggregating, comparing, and testing data to answer the Ask-phase question. Core operations are filter, sort, group, and calculate.
In spreadsheets, master SUM, AVERAGE, MIN, MAX, COUNT, COUNTA, COUNTIF, SUMIF, AVERAGEIF, plus pivot tables for cross-tabulation and VLOOKUP/XLOOKUP for lookups.
In SQL, know the join types: INNER (only matching rows), LEFT (all left rows plus matches), RIGHT (all right rows plus matches), and FULL OUTER (all rows from both sides, nulls where no match). Aggregate with GROUP BY and filter groups with HAVING; remember WHERE filters rows before aggregation while HAVING filters after. Use COUNT(DISTINCT column) for unique counts and subqueries for staged logic.
Statistically, distinguish correlation from causation, know that a confidence interval expresses uncertainty around an estimate, and understand that statistical significance depends on sample size as well as effect size. Validate results by cross-checking totals against a known source, spot-checking rows, and sanity-checking magnitudes.
Traps
The classic trap is WHERE versus HAVING; another is COUNT versus COUNT DISTINCT. Questions also present a rising trend with a request to conclude causation - the correct answer is that correlation alone does not establish it.
Rule of thumb
Write the question as a sentence first: group by what, filter which rows, compute which metric, compare to what baseline. Then choose the join and clause that matches that sentence.