Changelog
Source:NEWS.md
rfm 0.5.1
CRAN release: 2026-11-15
Bug fixes
- Corrected the customer-scoring return documentation and segment-summary plot title documentation.
- The package now requires dplyr 1.1.0 or later, matching the imported join API.
- Financial and fractional median plot labels now retain two decimal places instead of rounding to whole numbers; customer and order counts keep integer labels.
- Segment summary charts now show negative revenue and average order value, including all-negative data, with room for their labels.
- Interactive median and scatter hover text now names the selected metrics correctly.
- Deprecated scatter and histogram wrappers now forward axis labels, titles and
print_plotcorrectly and name their existing replacements. - Segment names such as
"1"are now assigned correctly, preserving first-match precedence and assigning unmatched customers to"Others". - Heatmaps now preserve an observed missing monetary mean as
NA, filling only absent score combinations with zero. - Automatic RFM bins now use right-closed quantile intervals without adding one to cut points, so currency conversions preserve monetary scores. Customers near a quantile boundary can receive different recency, frequency or monetary scores and segments; custom thresholds keep their lower-inclusive rule.
- Transaction scoring now keeps only metadata that is constant within each customer and names dropped columns in a message, preserving customer totals.
- Customer scoring now rejects duplicated customer IDs instead of multiplying rows and totals.
-
rfm_table_customer()now accepts either a date or a day-count recency column, with or without arecency_dayscolumn present (#111). -
rfm_table_customer_2()now runs instead of failing because of a missing argument (#112). -
rfm_table_order()andrfm_table_customer()now support different bin counts for recency, frequency and monetary; the threshold table is padded withNAto the longest dimension (#113). Segment definitions must fit the bin counts, as scores above the number of bins for a dimension never occur. -
rfm_heatmap_data()andrfm_plot_heatmap()now work with custom threshold vectors, and missing recency levels are no longer placed in the frequency column (#114).
rfm 0.5.0
CRAN release: 2026-10-16
This is a maintenance and trust restoration release.
Lifecycle Notice
-
rfmis now in maintenance mode: it will keep receiving fixes, and a successor package is in development.
Trust & CRAN Compliance Fixes
- Completely removed unsolicited network checks (
available.packages()), update prompts, andrunif()lottery from.onAttach(). Package attachment is now 100% offline and instantaneous. - Fixed report template corruption in
rfm_create_report(): changedfile.rename()to safefile.copy()(copy-only-if-missing), preserving user modifications and leaving the package installation template intact. - Eliminated all runtime interactive
install.packages()calls inrfm_create_report()andrfm_launch_app(). Functions now throw polite, structured, and informative errors viacli::cli_abort()in both interactive and non-interactive sessions. - Promoted
clitoImports. - Trimmed
Importsby demoting heavy visualization packages (plotly,gganimate,treemapify, andxplorerr) toSuggests, addinggifskitoSuggests. - Added runtime guards at all public entrypoints for optional interactive and animated features.
- Cleaned legacy test graphics artifacts (
tests/figs/,tests/testthat/Rplots.pdf).
rfm 0.3.0
CRAN release: 2024-02-26
Enhancements
- segment overview (#76)
-
rfm_create_report()generates a segmentation analysis report -
rfm_table_customer()handles customer level data with both recency date and days -
interactiveargument in plotting function generates interactive plots using plotly -
animateargument in plotting functions adds animation using gganimate
rfm 0.2.0
CRAN release: 2019-01-14
This is a minor release for bug fixes and enhancements.