Spotting address changes at a glance with jsdiffr

A tiny HTML diff that makes data cleaning reviewable

Note

This packages uses the jsdiffr package. Visit repo →

If you’ve ever standardised a column of addresses — expanding St to Street, fixing a postcode, normalising Ave — you know the hard part isn’t the transformation. It’s reviewing it. A before/after table of 5,000 rows is unreadable; your eye can’t find the one character that changed.

jsdiffr ships two HTML helpers, diff_to_html() and diff_html(), that turn a diff into colour-coded markup: red strike-through for what was removed, green for what was added. It’s a one-liner, and it works inline in Quarto, R Markdown, or a Shiny app.

One pair, one line

old <- "12 Baker St, London, NW1 6XE"
new <- "12 Baker Street, London, NW1 6EX"

cat(diff_to_html(diff_words(old, new)))

12 Baker StStreet, London, NW1 6XE6EX

At a glance: St became Street, and the postcode swapped its last two characters — a typo you’d never catch scanning two text columns.

A whole column, side by side

diff_html() takes two vectors and renders an Original / Updated / Changes table. Pass view = FALSE to get the HTML back instead of opening a browser.

original <- c(
  "350 Fifth Avenue, New York, NY 10118",
  "1600 Pennsylvania Ave NW, Washington DC",
  "Flat 4, 22 Acacia Avenue, Manchester M14 5TP"
)
cleaned <- c(
  "350 5th Avenue, New York, NY 10118",
  "1600 Pennsylvania Avenue NW, Washington, DC 20500",
  "Flat 4B, 22 Acacia Ave, Manchester M14 5TP"
)

cat(diff_table(original, cleaned, method = diff_words))

Left	Right	Changes
350 Fifth Avenue, New York, NY 10118	350 5th Avenue, New York, NY 10118	350 Fifth5th Avenue, New York, NY 10118
1600 Pennsylvania Ave NW, WashingtonDC	1600 Pennsylvania Avenue NW, Washington, DC 20500	1600 Pennsylvania AveAvenue NW, Washington, DC 20500
Flat 4, 22 Acacia Avenue, Manchester M14 5TP	Flat 4B, 22 Acacia Ave, Manchester M14 5TP	Flat 44B, 22 Acacia AvenueAve, Manchester M14 5TP

Every row tells you exactly what your cleaning step did — and, just as importantly, what it didn’t touch.

Hide the noise you don’t care about

Reviewing a St → Street expansion across thousands of rows? That’s expected; you don’t want it lighting up red and green. ignore_words renders changes made up entirely of those words as plain context, so only the surprising edits stand out:

cat(diff_table(
  original, cleaned,
  method       = diff_words,
  ignore_words = c("ave", "avenue", "5th", "fifth")
))

Left	Right	Changes
350 Fifth Avenue, New York, NY 10118	350 5th Avenue, New York, NY 10118	350 5th Avenue, New York, NY 10118
1600 Pennsylvania Ave NW, WashingtonDC	1600 Pennsylvania Avenue NW, Washington, DC 20500	1600 Pennsylvania Avenue NW, Washington, DC 20500
Flat 4, 22 Acacia Avenue, Manchester M14 5TP	Flat 4B, 22 Acacia Ave, Manchester M14 5TP	Flat 44B, 22 Acacia Ave, Manchester M14 5TP

Now the abbreviation expansions fade into the background and the genuine changes — the added ZIP code, 4 → 4B — are what catch your eye.

Why it’s worth reaching for

Reviewable cleaning. Diffs make an address-standardisation pass auditable instead of a black box.
Zero setup. No JS toolchain — diff_to_html() returns a string you can drop straight into a report or shiny::renderUI().
Tunable signal. ignore_words and the diff_* granularity (diff_chars, diff_words, diff_sentences) let you dial in exactly which changes matter.

Two functions, a couple of vectors, and a column of addresses becomes something you can actually read.