NEWS.md

In dataSDA: Datasets and Basic Statistics for Symbolic Data Analysis

dataSDA 0.2.6

New features

• check_zero_width_intervals() checks interval-valued data for zero-width intervals (min == max). It accepts both MM format (a data.frame with paired _min/_max columns) and RSDA format (a symbolic_tbl with symbolic_interval columns; non-interval columns are ignored). It returns (invisibly) a logical scalar carrying a "flagged" [n, p] matrix of degenerate cells and a "variables" vector of affected variables, and warns by default when any are found. Useful for screening data before passing it to downstream tools that divide by interval width (e.g. ggInterval::ggInterval_indexImage()). A tol argument allows near-zero widths to be flagged.

• aggregate_to_symbolic(type = "int") gains a zero_width argument controlling how zero-width intervals (min == max) in the aggregated output are handled: "keep" (default) leaves the output untouched, returning any zero-width intervals as-is without warning; "remove" drops every concept containing a zero-width interval; "regenerate" re-runs the aggregation (re-clustering or re-sampling) until none remain (effective only for stochastic group_by such as "kmeans"/"resampling"); and "adjust" adds a small epsilon (default 1e-07) to the upper endpoint of each zero-width interval. A companion epsilon argument sets the adjustment amount. For "remove" and "adjust", a single warning names the affected variables and the action taken. Such degenerate intervals otherwise break downstream tools such as ggInterval::ggInterval_indexImage() (which divides by interval width); use check_zero_width_intervals() to screen for them.

dataSDA 0.2.5

New datasets (9 added, 114 total)

Nine interval time series (ITS) datasets from global financial markets and meteorology:

Interval time series (9)

• sp500.its — S&P 500 index daily OHLC intervals.
• djia.its — Dow Jones Industrial Average daily OHLC intervals.
• ibovespa.its — Ibovespa (Brazil) daily OHLC intervals.
• crude_oil_wti.its — WTI crude oil daily OHLC intervals.
• merval.its — MERVAL (Argentina) daily OHLC intervals.
• petrobras.its — Petrobras stock daily OHLC intervals.
• euro_usd.its — EUR/USD exchange rate daily OHLC intervals.
• shanghai_stock.its — Shanghai Composite daily OHLC intervals.
• irish_wind.its — Irish wind energy daily intervals.

Dataset renaming

Renamed 7 datasets with proper type suffixes for consistency:

| Old name | New name | |---|---| | airline_flights2 | airline_flights2.modal | | crime | crime.modal | | crime2 | crime2.modal | | fuel_consumption | fuel_consumption.modal | | health_insurance2 | health_insurance2.modal | | occupations | occupations.modal | | occupations2 | occupations2.modal |

Also added mushroom.int.mm (interval multi-modal format of mushroom data).

New functions (8 exported)

• Data conversion: aggregate_to_symbolic() — convert traditional data to symbolic data format.
• I/O: read_symbolic_csv(), write_symbolic_csv() — read/write symbolic data as CSV.
• Search: search_data() — search available datasets by keyword or type.
• Format conversion: to_all_interval_formats() — convert intervals to all supported formats at once.
• ARRAY format converters: ARRAY_to_MM(), ARRAY_to_RSDA(), ARRAY_to_iGAP(), MM_to_ARRAY(), RSDA_to_ARRAY(), SODAS_to_ARRAY(), iGAP_to_ARRAY().

Bug fixes

• Fixed int_dist p parameter bug.
• Fixed documentation for 19 datasets with incorrect column metadata.
• Fixed function documentation: @usage defaults, @details, and examples.

Other changes

• Updated vignette with new examples and content.
• Pre-built vignette to avoid CRAN timeout (eval gated by NOT_CRAN).
• Updated all dataset documentation with @section Metadata blocks.
• R CMD check --as-cran: 0 errors, 0 warnings, 0 notes.

dataSDA 0.2.2

New datasets (17 added, 105 total)

Seventeen new datasets from R packages (intkrige, RSDA, MAINT.Data, GPCSIV, HistDAWass, symbolicDA), the Billard & Diday (2007) textbook, and the QualAr Portuguese air quality network:

Interval-valued (12)

• utsnow.int — 415 Utah weather stations with snow load prediction intervals plus coordinates/elevation (from intkrige).
• lynne1.int — 10 observations with pulse rate, systolic, and diastolic pressure intervals (from RSDA).
• loans_by_risk_quantile.int — 35 Lending Club loan groups (A1-G5) with 4 quantile-based financial intervals (from MAINT.Data).
• judge1.int — 6 regions rated by Judge 1 on 4 interval variables (from GPCSIV).
• judge2.int — 6 regions rated by Judge 2 on 4 interval variables (from GPCSIV).
• judge3.int — 6 regions rated by Judge 3 on 4 interval variables (from GPCSIV).
• video1.int — 10 user groups with 5 video engagement interval metrics (from GPCSIV).
• video2.int — 10 user groups with 5 video engagement interval metrics (from GPCSIV).
• video3.int — 10 user groups with 5 video engagement interval metrics (from GPCSIV).
• lisbon_air_quality.int — 1096 daily observations of 8 pollutant concentration intervals from Lisbon (QualAr).

Histogram-valued (4)

• blood.hist — 14 gender-age groups with cholesterol, hemoglobin, and hematocrit histograms (from HistDAWass).
• china_climate_month.hist — 60 Chinese weather stations with 168 monthly climate histograms (from HistDAWass).
• china_climate_season.hist — 60 Chinese stations with 56 seasonal climate histograms (from HistDAWass).
• exchange_rate_returns.hist — 108 monthly exchange rate return histograms (from HistDAWass).

Mixed symbolic (1)

• polish_cars.mix — 30 Polish car models with 9 interval + 3 multinomial variables (from symbolicDA).

From Billard & Diday (2007) textbook (2)

• hierarchy.hist — 10 observations with hierarchical categories, conditional histograms, and cholesterol interval (Table 6.20).
• bird_color_taxonomy.hist — 20 birds with density/size histograms, tone, and fuzzy shade taxonomy (Tables 6.9/6.14).

dataSDA 0.2.1

New datasets (19 added, 88 total)

Nineteen new datasets from Billard & Diday (2020) Clustering Methodology for Symbolic Data, the HistDAWass R package, and other R packages:

Interval-valued (5)

• genome_abundances.int — 14 genome classes with 10 dinucleotide abundance intervals (Table 3-16).
• china_temp_monthly.int — 15 Chinese weather stations with 12 monthly temperature intervals + elevation (Table 7-9).
• ecoli_routes.int — 9 E. coli transport routes with 5 biochemical interval variables (Table 8-10).
• loans_by_risk.int — 35 Lending Club loan groups by risk level (A-G) with 4 financial intervals (from MAINT.Data).
• polish_voivodships.int — 18 Polish voivodships with 9 socio-economic interval variables (from clusterSim).

Histogram-valued (9)

• iris_species.hist — 3 iris species with 4 morphological histogram variables (Table 4-10).
• flights_detail.hist — 16 airlines with 5 flight performance histograms (Table 5-1).
• cover_types.hist — 7 forest cover types with 4 topographic histograms (Table 7-21).
• glucose.hist — 4 regions with blood glucose histograms (Table 4-14).
• state_income.hist — 6 US states with 4 income distribution histograms (Table 7-18).
• simulated.hist — 5 simulated observations with 2 histogram variables (Table 7-26).
• age_pyramids.hist — 229 countries with 3 age pyramid histograms (from HistDAWass).
• ozone.hist — 84 daily observations with 4 weather histograms (from HistDAWass).
• french_agriculture.hist — 22 French regions with 4 agricultural histograms (from HistDAWass).

Distribution-valued (1)

• household_characteristics.distr — 12 counties with 3 categorical distribution variables (Table 6-1).

Mixed symbolic (4)

• county_income_gender.hist — 12 counties with gendered income histograms + sample sizes (Table 6-16).
• joggers.mix — 10 jogger groups with pulse rate intervals + running time histograms (Table 2-5).
• census.mix — 10 census regions with 6 mixed-type variables: histograms, distributions, multi-valued sets, and intervals (Table 7-23).
• mtcars.mix — 5 car groups with 7 interval + 4 modal variables (from ggInterval).

dataSDA 0.2.0

New datasets (13 added, 69 total)

Thirteen new datasets extracted from R packages and the Billard & Diday (2006) textbook:

From R packages

• cardiological.int — 44 patients with 5 interval-valued physiological measurements (from RSDA).
• prostate.int — 97 prostate cancer patients with 9 clinical interval variables (from RSDA).
• uscrime.int — 46 US states with 102 interval-valued crime statistics (from RSDA).
• hardwood.hist — 5 hardwood tree species with 4 histogram-valued climate variables (from RSDA).
• synthetic_clusters.int — 125 observations in 5 clusters with 6 interval variables (from symbolicDA).
• environment.mix — 14 EPA state groups with mixed interval/modal environmental data (from ggInterval).

From Billard & Diday (2006) textbook tables

• weight_age.hist — 7 age groups with histogram-valued weight distributions (Table 3.10).
• hospital.hist — 15 hospitals with histogram-valued cost distributions (Table 3.12).
• cholesterol.hist — 14 gender-age groups with cholesterol histograms (Table 4.5).
• hemoglobin.hist — 14 gender-age groups with hemoglobin histograms (Table 4.6).
• hematocrit.hist — 14 gender-age groups with hematocrit histograms (Table 4.14).
• hematocrit_hemoglobin.hist — 10 observations with bivariate 2-bin histograms (Table 6.8).
• energy_usage.distr — 10 towns with categorical fuel/heating distributions (Table 3.7).

dataSDA 0.1.9

New datasets (7 added, 55 total)

Seven new interval-valued benchmark datasets from recent SDA papers (2020-2025):

• freshwater_fish.int — 12 freshwater fish species with 13 heavy metal bioaccumulation variables, 4 feeding classes (Andrade et al., 2025).
• fungi.int — 55 fungi specimens with 5 morphological variables, 3 genera: Amanita, Agaricus, Boletus (Andrade et al., 2025).
• iris.int — 30 interval observations of Fisher's iris data, 4 sepal/petal variables, 3 species (Andrade et al., 2025).
• water_flow.int — 316 water flow sensor readings with 47 interval features, 2 classes (Andrade et al., 2025).
• wine.int — 33 wine samples with 9 chemical/physical property variables, 2 classes (Andrade et al., 2025).
• car_models.int — 33 Italian car models with 8 specification variables, 4 categories (Andrade et al., 2025).
• hdi_gender.int — 183 countries with 2 World Bank gender indicator intervals and ordinal HDI classification (Alcacer et al., 2023).

dataSDA 0.1.8

Vignette

• Comprehensive rewrite covering all 51 exported functions (was 18), organized into 14 sections: format detection/conversion, core stats, geometry, position, robust, shape, similarity, uncertainty, distance, histogram stats, and utilities.
• Fixed garbled characters, stale dataset references, and broken histogram example code.

dataSDA 0.1.7

New functions (2 exported)

• MM_to_RSDA() — convert MM format (_min/_max columns) to RSDA format (symbolic_tbl with complex-encoded intervals).
• iGAP_to_RSDA() — convert iGAP format to RSDA format via iGAP_to_MM → MM_to_RSDA.

Updated functions

• int_list_conversions() now returns 8 conversions (was 6), including MM_to_RSDA and iGAP_to_RSDA.
• int_convert_format() now supports to = "RSDA" for MM and iGAP sources with auto-detection.

Vignette

• Comprehensive rewrite covering all 51 exported functions (was 18), organized into 14 sections: format detection/conversion, core stats, geometry, position, robust, shape, similarity, uncertainty, distance, histogram stats, and utilities.
• Fixed garbled characters, stale dataset references, and broken histogram example code.

Tests

• 49 new tests for the new conversion functions and updated conversion registry (517 total, all passing).

dataSDA 0.1.6

Dataset naming convention

Adopted snake_case naming with type suffixes (.int, .hist, .mix, .distr, .iGAP) for all datasets. Renamed 10 existing datasets:

| Old name | New name | |---|---| | Abalone | abalone.int | | Abalone.iGAP | abalone.iGAP | | Cars.int | cars.int | | ChinaTemp.int | china_temp.int | | Face.iGAP | face.iGAP | | LoansbyPurpose.int | loans_by_purpose.int | | bird.int | bird.mix | | soccer.bivar.int | soccer_bivar.int | | airline_flights | airline_flights.hist | | health_insurance | health_insurance.mix |

New datasets (16 added, 48 total)

• acid_rain.int, bats.int, credit_card.int, employment.int, oils.int, teams.int, tennis.int, temperature_city.int, trivial_intervals.int, world_cup.int (interval-valued)
• bird_species.mix, bird_species_extended.mix, town_services.mix (mixed symbolic)
• lung_cancer.hist (histogram-valued)
• energy_consumption.distr (distribution-valued)
• bank_rates, mushroom_fuzzy (other)

New functions (3 exported)

• Format detection and conversion (interval_format_conversions.R):
• int_detect_format() — automatically detect interval data format (RSDA, MM, iGAP, SODAS).
• int_list_conversions() — list available format conversion functions, with optional filtering by source/target format.
• int_convert_format() — unified interface for all interval format conversions with auto-detection.

Other changes

• Updated vignette, examples, and tests to use new dataset names.
• 70 new tests for format detection and conversion helpers (468 total, all passing).
• R CMD check: 0 errors, 0 warnings, 0 notes.

dataSDA 0.1.5

New functions (31 exported)

• Interval distances (interval_dist.R): int_dist(), int_dist_matrix(), int_pairwise_dist(), int_dist_all() — 14 distance measures (GD, IY, L1, L2, CB, HD, EHD, nEHD, snEHD, TD, WD, minkowski, ichino, de_carvalho) with method aliases (euclidean, hausdorff, manhattan, city_block, wasserstein).
• Interval geometry (interval_geometry.R): int_width(), int_radius(), int_center(), int_overlap(), int_containment(), int_midrange().
• Position and scale (interval_position.R): int_median(), int_quantile(), int_range(), int_iqr(), int_mad(), int_mode() — all support 8 methods (CM, VM, QM, SE, FV, EJD, GQ, SPT).
• Robust statistics (interval_robust.R): int_trimmed_mean(), int_winsorized_mean(), int_trimmed_var(), int_winsorized_var().
• Distribution shape (interval_shape.R): int_skewness(), int_kurtosis(), int_symmetry(), int_tailedness().
• Similarity measures (interval_similarity.R): int_jaccard(), int_dice(), int_cosine(), int_overlap_coefficient(), int_tanimoto(), int_similarity_matrix().
• Uncertainty and variability (interval_uncertainty.R): int_entropy(), int_cv(), int_dispersion(), int_imprecision(), int_granularity(), int_uniformity(), int_information_content().

Refactoring

• Merged 6 interval conversion files into a single R/interval_format_conversions.R, organized by target format.
• Extracted shared internal helpers, reducing ~454 lines of duplicated code.
• R/interval_utils.R: consolidated 7 internal interval helpers, format-preparation functions (RSDA_format, set_variable_format, clean_colnames), and their internal helpers.
• R/histogram_utils.R: 8 internal helpers for histogram statistics.
• Renamed utils_validation.R to validation.R.
• R CMD check: 0 errors, 0 warnings, 0 notes. All 399 tests pass.

dataSDA 0.1.4

• Input validation for all 18 exported functions: every function now validates its inputs at entry, producing clear error messages instead of cryptic R internals errors.
• New R/utils_validation.R: 11 internal validation helpers centralizing all checks.
• RSDA_format fix: replaced 4 return("Error") with proper stop() calls.
• 62 new regression tests for input validation (399 total, all passing).
• R CMD check: 0 errors, 0 warnings.
• Added NEWS.md with changelog for all versions.

dataSDA 0.1.3

• Added testthat framework with 337 tests covering all 18 exported functions.
• 0 failures, 0 warnings, 0 skips.

dataSDA 0.1.2

• 18 exported functions for symbolic data format conversion and statistics.
• 32 datasets (interval-valued and histogram-valued) for symbolic data analysis.
• Support for MM, iGAP, RSDA, and SODAS data formats.
• Interval statistics: int_mean, int_var, int_cov, int_cor (8 methods).
• Histogram statistics: hist_mean, hist_var, hist_cov, hist_cor.

dataSDA documentation built on June 12, 2026, 9:06 a.m.