@karinadatascientist: Python statistics cheatsheet. Code on the left, how to read it on the right. Descriptive statistics. Mean, median, std, and a z-score formula for outliers, |z| > 3 is worth a look. Distributions. Shapiro-Wilk for normality (p > 0.05 looks normal), skew and kurtosis for shape. If your data isn't normal, that's your cue to transform it or switch tests. Hypothesis testing. T-test for two normal groups. Mann-Whitney U when they're not normal. ANOVA for three or more groups, tells you something differs, not what. Chi-square for categorical variables. A p-value tells you a result is real. It doesn't tell you if it's big enough to matter. Effect size. Cohen's d answers that: 0.2 small, 0.5 medium, 0.8 large. The number the original cheatsheet told you to check and never showed. Correlation and regression. Pearson for strength and direction. Linear regression for coefficients and R². Save this before your next analysis.

Karina.DataScientist

Open In TikTok:

Region: AE

Sunday 20 September 2026 16:51:46 GMT