Synthetic Data
I have spent a lot of time thinking and writing about synthetic data. It is the subject of my PhD.
These posts cover how synthetic data is generated, how to measure whether it is any good, and the trade-offs between privacy and utility that nobody escapes. I write from health research but the same problems show up in finance, pharma, and anywhere the real data is sensitive.
Bias in Synthetic Data
An exploration of bias in synthetic data and its implications for health research.
How Private is Synthetic Data? Understanding the Tradeoff with Utility
Synthetic data is a powerful tool for health research, but it comes with a tradeoff between privacy and utility. This blog explores what this means for researchers and how to navigate the tradeoff.
How Do We Measure the Utility of Synthetic Data?
A practical guide to some of the metrics you can use to evaluate the utility of synthetic data.
Why Synthetic Data is Good for Open Science
Understanding the benefits of synthetic data for open science and reproducibility.
Is your Synthetic Data actually private?
A practical guide to the three privacy risks in synthetic data, the metrics that quantify them, and why no single number tells you whether your data is safe.
Synthetic Data: The Complete Series
The jumping off point for my synthetic data series, covering the basics of synthetic data generation and its applications.
Representativeness in Synthetic Data: What It Means and How to Measure It
Understanding the concept of representativeness in synthetic data and the methods used to measure it.
How Synthetic Data Is Used in Healthcare, Research and Beyond
Explore real-world use cases for synthetic data in healthcare, clinical trials, finance and more.
Multiple Imputation and Perturbation: Why They're Not Built for Synthetic Data
This blog explores why multiple imputation and perturbation are not suitable for generating synthetic data.