Just enough statistics to be dangerous (and correct)

• Distributions & summary stats — mean vs. median, spread, skew.
• Correlation ≠ causation — the single most important idea in analytics.
• Sampling & uncertainty — your data is a sample; conclusions have error bars.
• Confidence intervals — a range, not a false-precision single number.

Why it matters

Statistics is what keeps you honest. It's the difference between "sales went up 3%" and "sales went up 3% ± 5%, so we can't yet tell." Decision-makers trust analysts who quantify their uncertainty.

Takeaway: you don't need heavy theory — you need the judgment to not fool yourself.