jaotusdiagramme

Jaotusdiagramme, or distribution diagrams, are a family of visualizations that depict how the values of a variable are distributed. They provide a compact overview of shape, central tendency, dispersion, and potential outliers in a data set. They are used to summarize univariate data and to compare distributions across groups.

Common forms include histograms, which show counts or frequencies in consecutive bins; density plots, which present a smoothed estimate of the probability density function; and cumulative distribution plots (CDFs), which display the probability that the variable is at most a given value. Box plots summarize the five-number distribution (minimum, first quartile, median, third quartile, maximum) and can highlight outliers. Violin plots combine density information with quartiles, while QQ plots compare a data distribution to a theoretical one.

Applications include exploring data characteristics, assessing normality, detecting skewness or multimodality, identifying outliers, and communicating results to a broad audience. They also support comparisons of distributions across treatment groups, time periods, or categories.

Construction considerations include the data type (usually univariate), bin width or bandwidth (in histograms and density plots, respectively), and axis scales (linear or logarithmic) to accommodate skewed data. CDFs can be empirical or theoretical, and box plots rely on quartile calculations.

Limitations arise from plotting choices and sample size: bin width or bandwidth can distort the apparent distribution, and small samples may yield misleading pictures. Proper labeling and context are essential for accurate interpretation.

See also: descriptive statistics, histograms, kernel density estimation, cumulative distribution function, box plot.