Making Fake Data Generators for Open Source Healthcare Data Science Projects
summary
Shows a working pipeline that produces fake tabular and image data for NHS projects and a separate synthetic route using synthvae trained on MIMIC‑III. Produces JSON‑driven generators (categorical options, nulls, ranges), uses sdv and a Kedro pipeline to run training and evaluation, and reports collision analysis, Gower distance, PCA and propensity tests. Code is going to be open sourced; synthetic outputs are not published yet.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:38NHS AI Lab Skunkworks
- 2:20Defining fake data
- 3:33Project case study
- 5:45Implementing fake data
- 8:05Best practices
- 9:56Considerations and examples
- 11:50Alternatives to fake data
- 13:59Synthetic data pipeline
- 16:24Evaluating synthetic data
- 19:29SDV and privacy checks
- 21:13Kedro pipeline tool
- 22:54Current status and future
- 23:40Q&A session
related talks
36mArchitecting Data: A Deep Dive Into the world of Synthetic Data
25mThe Importance of Synthetic Data in Data-Centric AI
19mIs this Real?: Generating Synthetic Data that Looks Real
17mOpen Source for Data Quality
45mSynthetic Data for Machine Learning Applications
46mPanel: Expanding the OpenSynth Ecosystem: Open Source Data and AI for Power Systems Modelling
43mBehind the Scenes at Gapminder: A Story about Data and Visualizations
47mDigital Decluttering