LoginSign Up
← Back to all sparks
D

dataSDA

INFRA · APIS
Velocity2.5

Datasets and Basic Statistics for Symbolic Data Analysis

github.com/hanmingwu1103/datasda ↗

dataSDA grew from a dataset collection into a symbolic-data conversion toolkit.

r-packagesymbolic-datainterval-datadataset-catalogueformat-conversion
◆Current state
The package now carries 105 documented datasets in interval, histogram, modal, and mixed symbolic formats, drawn from other R packages, the Billard and Diday textbooks, and public sources such as the Portuguese air quality network. Alongside the data it has accumulated conversion functions between the MM, RSDA, iGAP, SODAS, and ARRAY representations, CSV read and write support, and a keyword search over the catalogue.
◆Where it's heading
The arc across this window runs from cataloguing to tooling. Early releases added datasets and then spent two consecutive releases fixing format documentation across all 105 of them. Later releases shift to functions: format converters, symbolic CSV I/O, and most recently a diagnostic that flags zero-width intervals before they reach tools that divide by interval width. That last addition is the clearest signal of intent — the package is starting to guard the analyses downstream of it, not just supply inputs.
◆Prediction
Expect further validation helpers in the mould of the zero-width check, since interval data has several degenerate shapes that break downstream methods silently.

◆Recent moves

  1. 9d ago

    Corrects dataset references and citations

    A one-line release correcting dataset references, citations, and sources. Provenance housekeeping for a package whose value rests on attributing data to the papers it came from.

    View source ↗
  2. 2mo ago

    Adds zero-width interval diagnostics for symbolic data

    Adds a diagnostic that flags zero-width intervals in both supported formats, plus an argument controlling how aggregation handles them. It moves the package from supplying interval data to screening it before methods that divide by width fail.

    View source ↗
  3. 5mo ago

    Adds symbolic format converters, CSV I/O, and 11 interval series

    The largest functional release in the window: converters between all supported symbolic formats, symbolic CSV read and write, keyword search over the catalogue, and 11 interval time series datasets. Datasets were also renamed with type suffixes, making the catalogue self-describing.

    View source ↗
  4. 5mo ago

    Fixes column metadata for 19 interval datasets

    Corrects format metadata for 19 datasets with wrong column names, counts, or dimensions. Completes the documentation pass begun in the previous release.

    View source ↗
  5. 5mo ago

    Completes format documentation across all 105 datasets

    Documents 29 previously under-described datasets and adds cross-references between related format pairs. Necessary groundwork for the search and conversion functions that arrive two releases later.

    View source ↗
  6. 5mo ago

    Adds 17 datasets from R packages and reference texts

    Seventeen datasets added from six R packages, a textbook, and an air quality network, bringing the total to 105. For a catalogue package this is the core work rather than an increment.

    View source ↗