pointplot

Pointplot is a type of data visualization used to display point estimates of a numerical variable across the levels of a categorical variable, with error bars indicating uncertainty. It is commonly employed to compare group means or other statistics while conveying variability.

In the seaborn library, pointplot is a plotting function that groups data by the x axis category (and optionally by hue), computes an estimator for each group (mean by default), and draws a marker at that value with an accompanying error bar. The length of the error bar reflects a chosen measure of uncertainty, such as a confidence interval or standard deviation. The function supports options for the order of categories, color palettes, and the ability to specify a different estimator (for example, median) as well as the number of bootstrap samples used to compute the interval. It can also connect successive points with lines to illustrate a trend across ordered categories.

Usage of pointplot centers on comparing central tendencies between groups and visualizing the associated uncertainty. It is particularly useful for simple, aggregated comparisons and for highlighting patterns across an ordered or hierarchical structure when combined with hue.

Limitations include the potential loss of distributional detail, especially with small sample sizes or highly skewed data. Alternatives such as bar plots, box plots, violin plots, or swarm plots may be more appropriate when information about the data distribution is important.