Dataset Generators
Everything else in the catalogue is a finished dataset: one file, generated once, ready to download. This category is the engine that made it. You get a parameterised simulation model instead of a static export, and it produces data whenever you ask, at whatever sample size, parameter range and random seed you give it.
That matters when a project needs a pipeline rather than a download. Regenerate a cohort under new assumptions. Sweep a parameter across its whole range to see where your model breaks. Produce a fresh cohort every time something retrains. A fixed CSV sits there; a generator keeps working.
- Reproducible. Every run is seeded and documented, so you can get the identical cohort back or vary it on purpose.
- Parameterised. Population size, noise level, disease prevalence, dosing regimen, whatever the underlying model exposes is yours to set.
- Source-available. The mechanistic model driving the data can be read. You are not buying a black box.
Buy from this category when one download will not be enough, and what you really need is the ability to keep making the data.