dataSDA: datasets and basic statistics for symbolic data analysis in R

Traditional datasets typically represent each variable with a single value per observation. However, as data volume and complexity continue to grow, representing variables through high-level descriptors such as intervals, histograms, and probability distributions, collectively referred to as symbolic data, has become increasingly valuable. These enriched representations retain information about distributional structure and variability, thereby enhancing both analytical depth and interpretability. This paper introduces dataSDA, an R package developed to curate symbolic datasets across diverse research domains and to support their reading, writing, conversion, and summarization. Building upon the frameworks of RSDA and HistDAWass, dataSDA extends their functionality by providing unified format conversion with automatic detection, aggregation of conventional (single-valued) data into symbolic form, and functions for computing interval distances, similarity measures, and descriptive statistics. The package currently hosts 114 benchmark datasets. A subset of these is used to illustrate clustering, classification, and regression analyses, as well as exploratory data analysis and visualization with the ggInterval package. By integrating ready-to-use symbolic datasets with tools for data format transformation and descriptive statistics, dataSDA aims to serve as a comprehensive resource for symbolic data collection and analysis. The package promotes accessibility, transparency, and reproducibility in symbolic data research and is freely available on CRAN.