UCLH-DHCT / emap

All core Emap functions (monorepo)
2 stars 1 forks source link

Synthetic generator should produce more realistic time/location data #50

Open jeremyestein opened 1 month ago

jeremyestein commented 1 month ago

If real data is not going to be up any time soon, in order to fully test the location finding functionality, the synthetic generator will need to be improved so it generates synethic data with location + time that match up to real location visits.

You could (and given time pressures, we probably will) just generate data for all ICU beds and assume that some of them will correspond to a real patient occupancy, but you will then end up with a lot of orphan data that could disguise bugs.

This will be slightly hard because the generator is quite upstream in the pipeline, and would not normally have access to Emap to know when genuine location visits existed.

jeremyestein commented 5 days ago

Note that also the observation times it generates are fixed times well before the present, so we don't get to test the live data scenario, where there may be more fragmentation due to data being a second or two delayed or out of order.