Data Frame Fingerprints and Lineage Figures

Profiles R data frames as compact data fingerprints using schema, shape, missingness, distribution, category, uniqueness, time, and role signals. It compares versions, identifies close relatives in a library of historical data sets, and renders portable HTML cards plus static PNG/PDF lineage figures for reports.

DataDNA

DataDNA is an R package that gives every data frame a compact fingerprint, lineage match, and report-ready identity figure.

Instead of only asking "what is in this table?", DataDNA asks:

• What kind of data set is this?
• How stable is its identity?
• Did this version drift from the previous one?
• Which columns changed their role, missingness, categories, or distribution?

The package is designed for analysts who receive CSVs, extracts, dashboards, or modeling data sets and need a fast way to recognize and compare them.

Example

library(DataDNA)

demo <- dna_example_customers()

dna <- data_dna(demo$customers_new, name = "customers_new")
dna

card <- dna_card(dna, file = "customers_dna.html")

dna_compare(demo$customers_old, demo$customers_new)
dna_diff(demo$customers_old, demo$customers_new)

dna_compare() combines exact schema overlap with shape, species, role structure, distribution, missingness, category, and identity signals. This makes the score feel more like a data fingerprint than a strict column-name check.

The package also includes lazy-loaded customers_old and customers_new example data sets.

Find the closest ancestor

library <- list(
  customers_2024 = data_dna(customers_old),
  customers_2025 = data_dna(customers_new)
)

match <- dna_match(customers_new, library)
match

dna_match_plot(match, file = "lineage.png")

dna_match_plot() is now the recommended reporting output. It renders a static PNG/PDF lineage figure with base R graphics: white background, compact ranking table, and restrained similarity lines that fit technical reports, papers, and slide decks better than a web page.

Core API

data_dna(df)
dna_card(df)
dna_compare(old_df, new_df)
dna_diff(old_df, new_df)
dna_match(new_df, dna_library)
dna_match_card(match)
dna_match_plot(match)
dna_species(df)

Installation

From GitHub:

install.packages("devtools")
devtools::install_github("TonyIsFool/DataDNA")

Or with the lighter remotes package:

install.packages("remotes")
remotes::install_github("TonyIsFool/DataDNA")

From a local source tarball:

install.packages("DataDNA_0.1.0.tar.gz", repos = NULL, type = "source")

Design

The profiling and comparison algorithms use base R. The HTML card uses the lightweight htmltools package so the result is portable and CRAN-friendly.