Intro to refinr

Chris Muir

Package Information

The purpose of this package is to assist in working with strings that are effectively equivalent, yet are not quite identical. The functions take a character vector as input, identify and cluster similar values, and then merge clusters together so their values become identical. The clustering performed by these functions are implementations of the “key collision” and “ngram fingerprint” algorithms from the open source tool Open Refine. More info on key collision and ngram fingerprint can be found here.

In addition, there are a few add-on features included, to make the clustering/merging functions more useful. These include approximate string matching to allow for merging despite minor mispellings, the option to pass a dictionary vector to dictate edit values, and the option to pass a vector of strings to ignore during the clustering process.

This package provides two exported functions, key_collision_merge and n_gram_merge. Below is an explanation of each.

key_collision_merge

library(refinr)
x <- c("Acme Pizza, Inc.", "AcMe PiZzA, Inc.", "ACME PIZZA COMPANY", "acme pizza LLC")
key_collision_merge(x)
#> [1] "ACME PIZZA COMPANY" "ACME PIZZA COMPANY" "ACME PIZZA COMPANY"
#> [4] "ACME PIZZA COMPANY"

Argument bus_suffix allows the clustering to be insensitive to common business name suffix strings (i.e. “inc”, “llc”, “co”, etc.). The default value is TRUE.

# Set bus_suffix to FALSE to see the difference (only the first two strings get merged).
key_collision_merge(x, bus_suffix = FALSE)
#> [1] "AcMe PiZzA, Inc."   "AcMe PiZzA, Inc."   "ACME PIZZA COMPANY"
#> [4] "acme pizza LLC"

A character vector can be passed to argument dict, which will dictate merge values when a cluster has a match within the dict vector.

key_collision_merge(x, dict = c("Acme Pizza, Incorporated"))
#> [1] "Acme Pizza, Incorporated" "Acme Pizza, Incorporated"
#> [3] "Acme Pizza, Incorporated" "Acme Pizza, Incorporated"

To specify strings to ignore during the clustering process, pass a character vector to argument ignore_strings.

x <- c("Bakersfield Highschool", "BAKERSFIELD high", "high school, bakersfield")
key_collision_merge(x, ignore_strings = c("high", "school", "highschool"))
#> [1] "BAKERSFIELD high" "BAKERSFIELD high" "BAKERSFIELD high"

These args can also be used in combination with each other.

key_collision_merge(x, ignore_strings = c("high", "school", "highschool"), dict = c("Bakersfield High School"))
#> [1] "Bakersfield High School" "Bakersfield High School"
#> [3] "Bakersfield High School"

n_gram_merge

Works similarly to key_collision_merge, however it features approximate string matching, which allows for merging of strings that contain slight spelling differences. Package stringdist is used for calculating edit distance between strings.

x <- c("Acme Pizza, Inc.", "ACME PIZA COMPANY", "Acme Pizzazza LLC")
n_gram_merge(x)
#> [1] "ACME PIZA COMPANY" "ACME PIZA COMPANY" "ACME PIZA COMPANY"

The performance of the approximate string matching can be ajusted using parameters weight and/or edit_threshold.

n_gram_merge(x, weight = c(d = 1, i = 0.4, s = 0.2, t = 0.2))
#> [1] "ACME PIZA COMPANY" "ACME PIZA COMPANY" "Acme Pizzazza LLC"

Additional arguments can also be supplied that will be passed along to function stringdistmatrix from the stringdist package.

n_gram_merge(x, method = "soundex", useBytes = TRUE)
#> [1] "ACME PIZA COMPANY" "ACME PIZA COMPANY" "Acme Pizzazza LLC"

n_gram_merge also features arguments bus_suffix and ignore_strings, that operate the same way they do for function key_collision_merge.

x <- c("Bakersfield Highschool", "BAKERSFIELD high", "high school, bakersfield")
n_gram_merge(x, ignore_strings = c("high", "school", "highschool"))
#> [1] "BAKERSFIELD high" "BAKERSFIELD high" "BAKERSFIELD high"

Example Workflow

library(dplyr)

x <- c("Clemsson University", 
       "university-of-clemson", 
       "CLEMSON", 
       "Clem son, U.", 
       "college, clemson u", 
       "M.I.T.", 
       "Technology, Massachusetts' Institute of", 
       "Massachusetts Inst of Technology", 
       "UNIVERSITY:  mit"
)

ignores <- c("university", "college", "u", "of", "institute", "inst")
x_refin <- x %>% 
  key_collision_merge(ignore_strings = ignores) %>% 
  n_gram_merge(ignore_strings = ignores)

# Print results.
cat(paste(x_refin, collapse = "<br />"))

CLEMSON
CLEMSON
CLEMSON
CLEMSON
CLEMSON
M.I.T.
Massachusetts Inst of Technology
Massachusetts Inst of Technology
M.I.T.

# Create df for comparing the original values to the edited values.
# This is especially useful for larger input vectors.
inspect_results <- data_frame(original_values = x, edited_values = x_refin) %>% 
  mutate(equal = original_values == edited_values)

# Display only the values that were edited by refinr.
knitr::kable(
  inspect_results[!inspect_results$equal, c("original_values", "edited_values")], 
  format = "html", 
  table.attr = "style='width:100%;'"
)
original_values edited_values
Clemsson University CLEMSON
university-of-clemson CLEMSON
Clem son, U. CLEMSON
college, clemson u CLEMSON
Technology, Massachusetts’ Institute of Massachusetts Inst of Technology
UNIVERSITY: mit M.I.T.