Descriptive statistics summarise every selected column at once: n, missing, mean, SD, SE, 95% CI of the mean, median, IQR, min/Q1/Q3/max, skewness and excess kurtosis.
Almost every analysis starts with a quick numerical sketch of each variable: how complete is the data, where is its centre, how spread out is it, is it skewed. summary_stats lays all of that out in a single table you can copy into a methods section or a Table 1.
Both the parametric (mean, SD) and the robust (median, IQR) summaries are shown side by side. When they disagree substantially (mean ≠ median, or SD ≫ IQR/1.349), the distribution is skewed or has heavy tails and the robust column is the better one to report.
One or more numeric columns.
One row per column with all summary statistics, including 95% CI for the mean (t-based).
Skewness near 0 with kurtosis near 0 ⇒ data are approximately normal. |Skewness| > 1 ⇒ noticeable asymmetry; consider a log/sqrt transform. Excess kurtosis > 3 ⇒ heavy tails; consider robust methods.
A large mean–median gap is the simplest signal of skew. Use the median + IQR pair to describe skewed data; reserve mean + SD for symmetric distributions.