Delighted to share our new paper detailing how to create synthetic data to help promote open-science practices (specifically within longitudinal population-based studies) using #ALSPAC as an example wellcomeopenresearch.org/articles/9-5... /1 - ThreadSky

djsmith90.bsky.social • 378 days ago

Delighted to share our new paper detailing how to create synthetic data to help promote open-science practices (specifically within longitudinal population-based studies) using #ALSPAC as an example

https://wellcomeopenresearch.org/articles/9-57/v1

/1

Comments

djsmith90.bsky.social•378 days ago

Why is this important?

Making data (and analysis code) open is vital for ensuring reproducibility, transparency and trustworthiness in science. But sometimes this isn’t possible, as data may be either sensitive, identifiable or have sharing restrictions in place

/2

djsmith90.bsky.social•378 days ago

This is the case for many epidemiological studies, especially those using longitudinal population-based datasets (which may require specific regulations for users). This includes research using ALSPAC data and other datasets (UK Biobank, for example)

/3

djsmith90.bsky.social•378 days ago

In this paper, we use the ‘synthpop’ R package (https://synthpop.org.uk) to generate synthetic datasets. These datasets are wholly synthetic – and therefore contain no disclosive individual-level data – while maintaining the features and relations of the original data

/4

djsmith90.bsky.social•378 days ago

We provide an easy-to-follow example using synthpop to generate synthetic ALSPAC data, with an openly-available subset of ALSPAC data (https://osf.io/8sgze). We also describe ALSPAC guidelines on how to make synthetic ALSPAC data publicly-available (for use in any study)

/5

djsmith90.bsky.social•378 days ago

This synthetic data approach provides a reasonable compromise between the competing demands of making datasets openly-available and allowing studies to maintain control over access to potentially-sensitive data. This approach allows research to be ‘quasi-reproducible’

/6

djsmith90.bsky.social•378 days ago

Researchers can then use these synthetic datasets to follow the analysis code, replicate the results in the paper (albeit not perfectly, as the data is not 100% identical), explore the methods used, and more generally check the analysis for any issues or errors

/7

Comments

Posting Rules

Reply