The R package fuzzylink implements a probabilistic record linkage
procedure proposed in Ornstein
(2025). This method allows
users to merge datasets with fuzzy matches on a key identifying
variable. Suppose, for example, you have the following two datasets:
dfA
#> name age
#> 1 Joe Biden 81
#> 2 Donald Trump 77
#> 3 Barack Obama 62
#> 4 George W. Bush 77
#> 5 Bill Clinton 77
dfB
#> name
#> 1 Joseph Robinette Biden
#> 2 Donald John Trump
#> 3 Barack Hussein Obama
#> 4 George Walker Bush
#> 5 William Jefferson Clinton
#> 6 George Herbert Walker Bush
#> 7 Biff Tannen
#> 8 Joe Riley
#> hobby
#> 1 Football
#> 2 Golf
#> 3 Basketball
#> 4 Reading
#> 5 Saxophone
#> 6 Skydiving
#> 7 Bullying
#> 8 JoggingWe would like a procedure that correctly identifies which records in
dfB are likely matches for each record in dfA. The fuzzylink()
function performs this record linkage with a single line of code.
library(fuzzylink)
df <- fuzzylink(dfA, dfB, by = 'name', record_type = 'person')
df
#> A
#> 1 Joe Biden
#> 2 Donald Trump
#> 3 Barack Obama
#> 4 George W. Bush
#> 5 Bill Clinton
#> B
#> 1 Joseph Robinette Biden
#> 2 Donald John Trump
#> 3 Barack Hussein Obama
#> 4 George Walker Bush
#> 5 William Jefferson Clinton
#> sim jw match
#> 1 0.7661285 0.7673401 Yes
#> 2 0.8388663 0.9333333 Yes
#> 3 0.8457284 0.9200000 Yes
#> 4 0.8445312 0.9301587 Yes
#> 5 0.8730800 0.5788889 Yes
#> match_probability age hobby
#> 1 1 81 Football
#> 2 1 77 Golf
#> 3 1 62 Basketball
#> 4 1 77 Reading
#> 5 1 77 Saxophone
The procedure works by using pretrained text embeddings to construct a measure of similarity for each pair of names. These similarity measures are then used as predictors in a statistical model to estimate the probability that two name pairs represent the same entity. See Ornstein (2025) for technical details.
You can install fuzzylink from CRAN with:
install.packages('fuzzylink')Or you can install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("joeornstein/fuzzylink")You will also need API access to a large language model (LLM). The
fuzzylink package currently supports OpenAI, Mistral, and Anthropic
Claude LLMs, but will default to using OpenAI unless specified by the
user.
Sign up for a developer account with OpenAI, then create an API key through your profile. For best performance, I strongly recommend purchasing at least $5 in API credits, which will significantly increase your API rate limits.
Once your account is created, copy-paste your API key into the following line of R code.
library(fuzzylink)
openai_api_key('YOUR API KEY GOES HERE', install = TRUE)
If you prefer to use language models from Mistral, you can sign up for an account here. As of writing, Mistral requires you to purchase prepaid credits before you can access their language models through the API.
Once you have a paid account, you can create an API key here, and copy-paste the API key into the following line of R code:
library(fuzzylink)
mistral_api_key('YOUR API KEY GOES HERE', install = TRUE)
If you prefer to use Anthropic’s Claude models, you can sign up for an account here. Once you have an API key, copy-paste it into the following line of R code:
library(fuzzylink)
anthropic_api_key('YOUR API KEY GOES HERE', install = TRUE)
Now you’re all set up!
Here is some code to reproduce the example above and make sure that everything is working on your computer.
library(tidyverse)
library(fuzzylink)
dfA <- tribble(~name, ~age,
'Joe Biden', 81,
'Donald Trump', 77,
'Barack Obama', 62,
'George W. Bush', 77,
'Bill Clinton', 77)
dfB <- tribble(~name, ~hobby,
'Joseph Robinette Biden', 'Football',
'Donald John Trump ', 'Golf',
'Barack Hussein Obama', 'Basketball',
'George Walker Bush', 'Reading',
'William Jefferson Clinton', 'Saxophone',
'George Herbert Walker Bush', 'Skydiving',
'Biff Tannen', 'Bullying',
'Joe Riley', 'Jogging')
df <- fuzzylink(dfA, dfB, by = 'name', record_type = 'person')
dfIf the df object links all the presidents to their correct name in
dfB, everything is running smoothly! (Note that you may see a warning
from glm.fit. This is normal. The stats package gets suspicious
whenever the model fit is too perfect.)
-
The
byargument specifies the name of the fuzzy matching variable that you want to use to link records. The dataframesdfAanddfBmust both have a column with this name. -
The
record_typeargument should be a singular noun describing the type of entity thebyvariable represents (e.g. “person”, “organization”, “interest group”, “city”). It is used as part of a language model prompt when training the statistical model. -
The
instructionsargument should be a string containing additional instructions to include in the language model prompt. Format these like you would format instructions to a human research assistant, including any relevant information that you think would help the model make accurate classifications. -
The
modelargument specifies which language model to prompt. It defaults to OpenAI’s'gpt-5.2', but also accepts Mistral models (e.g.'mistral-large-latest') and Anthropic Claude models (e.g.'claude-sonnet-4-5-20250929'). For simpler problems, you can try'gpt-3.5-turbo-instruct', which will significantly reduce cost and runtime. -
The
embedding_modelargument specifies which pretrained text embeddings to use when modeling match probability. It defaults to OpenAI’s ‘text-embedding-3-large’, but will also accept ‘text-embedding-3-small’ or Mistral’s ‘mistral-embed’. -
Several parameters—including
p,k,embedding_dimensions,max_validations, andparallel—are for advanced users who wish to customize the behavior of the algorithm. See the package documentation for more details. -
If there are any variables that must match exactly in order to link two records, you will want to include them in the
blocking.variablesargument. As a practical matter, I strongly recommend including blocking variables wherever possible, as they reduce the time and cost necessary to compute pairwise distance metrics. Suppose, for example, that our two illustrative datasets have a column calledstate, and we want to instructfuzzylink()to only link people who live within the same state.
dfA <- tribble(~name, ~state, ~age,
'Joe Biden', 'Delaware', 81,
'Donald Trump', 'New York', 77,
'Barack Obama', 'Illinois', 62,
'George W. Bush', 'Texas', 77,
'Bill Clinton', 'Arkansas', 77)
dfB <- tribble(~name, ~state, ~hobby,
'Joseph Robinette Biden', 'Delaware', 'Football',
'Donald John Trump ', 'Florida', 'Golf',
'Barack Hussein Obama', 'Illinois', 'Basketball',
'George Walker Bush', 'Texas', 'Reading',
'William Jefferson Clinton', 'Arkansas', 'Saxophone',
'George Herbert Walker Bush', 'Texas', 'Skydiving',
'Biff Tannen', 'California', 'Bullying',
'Joe Riley', 'South Carolina', 'Jogging')df <- fuzzylink(dfA, dfB,
by = 'name',
blocking.variables = 'state',
record_type = 'person')
df#> A
#> 1 Joe Biden
#> 2 Barack Obama
#> 3 George W. Bush
#> 4 Bill Clinton
#> 5 Donald Trump
#> B
#> 1 Joseph Robinette Biden
#> 2 Barack Hussein Obama
#> 3 George Walker Bush
#> 4 William Jefferson Clinton
#> 5 <NA>
#> sim block jw match
#> 1 0.7661773 1 0.7673401 Yes
#> 2 0.8457001 3 0.9200000 Yes
#> 3 0.8446921 4 0.9301587 Yes
#> 4 0.8732265 5 0.5788889 Yes
#> 5 NA NA NA <NA>
#> match_probability state age
#> 1 1 Delaware 81
#> 2 1 Illinois 62
#> 3 1 Texas 77
#> 4 1 Arkansas 77
#> 5 NA New York 77
#> hobby
#> 1 Football
#> 2 Basketball
#> 3 Reading
#> 4 Saxophone
#> 5 <NA>
Note that because Donald Trump is listed under two different states—New
York in dfA and Florida in dfB–the fuzzylink() function no longer
returns a match for this record; all blocking variables must match
exactly before the function will link two records together. You can
specify as many blocking variables as needed by inputting their column
names as a vector.
The function returns a few additional columns along with the merged
dataframe. The column match_probability reports the model’s estimated
probability that the pair of records refer to the same entity. This
column should be used to aid in validation and can be used for computing
weighted averages if a record in dfA is matched to multiple records in
dFB. The columns sim and jw are string distance measures that the
model uses to predict whether two records are a match. And if you
included blocking.variables in the function call, there will be a
column called block with an ID variable denoting which block the
records belong to.
Because the fuzzylink() function makes several calls to the OpenAI
API—which charges a per-token fee—there is a monetary cost associated
with each use. Based on the package defaults and API pricing as of
February 2026, here is a table of approximate costs for merging datasets
of various sizes.
| dfA | dfB | Approximate Cost (Default Settings) |
|---|---|---|
| 10 | 10 | $0 |
| 10 | 100 | $0 |
| 10 | 1,000 | $0 |
| 10 | 10,000 | $0.01 |
| 10 | 100,000 | $0.06 |
| 10 | 1,000,000 | $0.59 |
| 100 | 10 | $0.03 |
| 100 | 100 | $0.03 |
| 100 | 1,000 | $0.03 |
| 100 | 10,000 | $0.03 |
| 100 | 100,000 | $0.08 |
| 100 | 1,000,000 | $0.61 |
| 1,000 | 10 | $0.26 |
| 1,000 | 100 | $0.26 |
| 1,000 | 1,000 | $0.26 |
| 1,000 | 10,000 | $0.27 |
| 1,000 | 100,000 | $0.32 |
| 1,000 | 1,000,000 | $0.85 |
| 10,000 | 10 | $0.53 |
| 10,000 | 100 | $0.53 |
| 10,000 | 1,000 | $0.53 |
| 10,000 | 10,000 | $0.54 |
| 10,000 | 100,000 | $0.59 |
| 10,000 | 1,000,000 | $1.12 |
| 100,000 | 10 | $0.58 |
| 100,000 | 100 | $0.58 |
| 100,000 | 1,000 | $0.58 |
| 100,000 | 10,000 | $0.59 |
| 100,000 | 100,000 | $0.64 |
| 100,000 | 1,000,000 | $1.17 |
| 1,000,000 | 10 | $1.11 |
| 1,000,000 | 100 | $1.11 |
| 1,000,000 | 1,000 | $1.11 |
| 1,000,000 | 10,000 | $1.12 |
| 1,000,000 | 100,000 | $1.17 |
| 1,000,000 | 1,000,000 | $1.69 |
Note that cost scales more quickly with the size of dfA than with
dfB, because it is more costly to complete LLM prompts for validation
than it is to retrieve embeddings. With particularly large datasets, one
can reduce costs by using GPT-3.5 (model = 'gpt-3.5-turbo-instruct'),
blocking (blocking.variables), or reducing the maximum number of pairs
labeled by the LLM (max_labels).