The Curious Engineer

All blogs / Databases

Robin Linacre's homepage: Probabilistic record linkage, Data Science a

robinlinacre.com

Probabilistic record linkage, Data Deduplication, Data Science, Engineering and the Environment

DatabasesPythonAI / MLSoftware / engineering practice

’18’19’20’21’22’23’24’25’26
Posts
73 from 5 Apr 2018 to 19 Oct 2026
Posting
monthly (13 in the last year)
Kind
personal blog
Feed
https://www.robinlinacre.com/rss.xml

2026 11

  1. Using AI agents effectively without trusting them
  2. Click here for probabilistic linkage training materials homepage.
  3. Microblog
  4. Quotes and Links
  5. Agile Price Forecast
  6. Data Science in the Age of AI
  7. letterpaths: or how LLMs can be good even when they're bad
  8. Optimising DuckDB performance on large EC2 instances
  9. Letterpaths - a free library for teaching cursive writing
  10. Splink or Swim: The Benefits of Being Small Enough To Fail
  11. Respectful use of AI in software development teams

2025 9

  1. Measuring the accuracy of record linkage
  2. Using a fault tolerant trie for address matching
  3. Building Accurate Address Matching Systems
  4. Putting Scaffolding Around Vibe Coding to Build More Complex Apps
  5. Why DuckDB is my first choice for data processing
  6. An alternative way to think about predicted probabilities in the Fellegi Sunter model
  7. Live DuckDB WASM Splink model
  8. Graph editor for illustrating clustering concepts (graph playground)
  9. AI probably won't replace me in 2025

2024 5

  1. The emerging impact of LLMs on my productivity
  2. Splink: Transforming data linking through open source collaboration
  3. Connected components visualisation
  4. Match weight calculator
  5. Super-fast deduplication of large datasets using Splink and DuckDB

2023 10

  1. Why Probabilistic Linkage is More Accurate than Fuzzy Matching For Data Deduplication
  2. Thoughts and questions about the short term impact of LLMs on knowledge workers
  3. Visualising updating a prior
  4. Computing the Fellegi Sunter model
  5. m and u values in the Fellegi-Sunter model
  6. Partial match weights
  7. The relationship between probabilities, match weights and Bayes factors
  8. Splink and the Open Source Dividend
  9. SQL should be the default choice for data transformation logic
  10. Why parquet files are my preferred API for bulk open data

2022 3

  1. Why don't you just
  2. The Intuition Behind the Use of Expectation Maximisation to Train Record Linkage Models
  3. Splink 3: Fast, accurate and scalable linkage in Python

2021 13

  1. m and u probability generator with starting values
  2. Are more complex probabilistic linkage models more accurate? Part 2, unsupervised learning
  3. Are more complex probabilistic linkage models more accurate? Part 1, supervised learning
  4. The Thorniest Problem of Building an Analytical Platform
  5. The carbon impact of switiching to an electric car
  6. m and u probability generator
  7. Dependencies between match weights
  8. Understanding match weights in the Fellegi Sunter model
  9. Visualising the Fellegi Sunter model
  10. Maths of Fellegi Sunter (old version)
  11. The mathematics of the Fellegi Sunter model
  12. Say thanks!
  13. An Interactive Introduction to Record Linkage (Data Deduplication) in the Fellegi-Sunter framework

2020 8

  1. The Downfall of Command and Control Data Leadership
  2. Demystifying Apache Arrow
  3. Birdsong quiz
  4. Birdsong recording finder
  5. Comparing energy usage across countries
  6. Filling the country with solar panels
  7. Fuzzy Matching and Deduplicating Hundreds of Millions of Records with Splink
  8. Why you should open source your analytical work

2019 11

  1. Understanding the Spark UI by example: sorting data
  2. Understanding the Spark UI by example: the Left Join
  3. Spark UI SQL detailed annotator
  4. Unsupervised probabalistic data matching using the Expectation Maximisation algorithm
  5. Carbon offsetting vs. the cost of renewable energy
  6. Interactive blogging with Observable Notebooks and gatsby.js
  7. Flight distance calculator
  8. Energy usage ready reckoner
  9. My flights
  10. Effective testing of analytical models using automated sense checks
  11. Questions Senior Leaders Should Ask Their Data Delivery Teams

2018 3

  1. Why I’m backing Vega-Lite as our default tool for data visualisation
  2. Transforming analytical functions by mainstreaming data science
  3. Pushing the boundaries of data science with the MOJ Analytical Platform