splink 0.1.5

[Beta]: Implementation in Apache Spark of the EM algorithm to estimate parameters of Fellegi-Sunter's canonical model of record linkage.

splink: Probabalistic record linkage at scale

WARNING: Splink is is currently in beta testing. Please feel free to try it, but note this software is not fully tested, and the interface is likely to continue to change.

splink implements Fellegi-Sunter's canonical model of record linkage in Apache Spark, including EM algorithm to estimate parameters of the model.

The aims of splink are to:

• Work at much greater scale than current open source implementations (100 million records +).

• Get results faster than current open source implementations - with runtimes of less than an hour.

• Have a highly transparent methodology, so the match scores can be easily explained both graphically and in words

• Have accuracy similar to some of the best alternatives

Installation

splink is a Python package. It uses the Spark Python API to execute data linking jobs in a Spark cluster. It has been tested in Apache Spark 2.3 and 2.4.

Install splink using

pip install splink

Interactive demo

You can run demos of splink in an interactive Jupyter notebook by clicking the button below:

Documentation

The best documentation is currently a series of demonstrations notebooks in the splink_demos repo.

We also provide an interactive splink settings editor and example settings here

The statistical model behind splink is the same as that used in the R fastLink package. Accompanying the fastLink package is an academic paper that describes this model. This is the best place to start for users wanting to understand the theory about how splink works.

Video intro

You can find a short video introducing splink and running though an introductory demo here.

Acknowledgements

We are grateful to ADR UK (Administrative Data Research UK) for providing funding for this work as part of the Data First project.