statistical-analysis

To choose the right hypothesis test for A/B testing, match the data type and distribution to methods like t-tests for means, chi-squared for proportions, or nonparametric tests for non-normal data. It provides guidance on interpreting p-values, effect sizes, and confidence intervals.

The best way to detect outliers in a time series is using robust methods like z-score, IQR, or percentile thresholds. It recommends appropriate estimators and moving average smoothing to distinguish genuine extremes from data errors in business datasets.

To calculate descriptive statistics for skewed distributions, use robust estimators like the median and IQR instead of the mean and standard deviation. It provides guidance on selecting percentiles to accurately characterize asymmetric data distributions.

Yes, you can use moving averages for simple time-series forecasting. It provides options for moving average smoothing, period-over-period comparisons, and naive or seasonal forecasts while communicating uncertainty in the trend analysis results.

You should use nonparametric tests instead of ANOVA when your data does not meet normality assumptions or has small sample sizes. It offers a decision framework to help you select the correct hypothesis test and interpret the resulting effect sizes confidently.