[ ] read from flow.ed_predictor_csn and store as json data on the page
[ ] ditto from flow.ed_predictor_agg at 8 hours (i.e. something like the length of the working shift)
[ ] add a paginated table of patients with their prob of admission
Not clear what would be best to show
Simple example and then leave Zella to work on this?
Maybe we can work with flow.ed_predictor_csn?
Would be nice to see if we can show this as a trend over time but will need adjusting for time of day and day of week
Data source as per note from Zella below
With regard to accessing the ED predictions, there are two tables on flow:
flow.ed_predictor_csn has all adult patients in ED at the time of the predictions, with their current location and their admission probability
flow.ed_predictor_agg has predictions over a number of beds for these patients. This table has multiple distributions on it (and with hindsight, should have been made to be much easier to interrogate). If you want a time window of four hours for all patients including those not yet arrived, set time_window = 4, dist = "Empirical poisson". Other values for time_window are 8 or (other than when the report is generated at 4 pm) the number of hours between the report time and midnight.
Sorry - this is probably really confusing. Please shout if you need clarification.
Initial notes
Not clear what would be best to show Simple example and then leave Zella to work on this? Maybe we can work with flow.ed_predictor_csn?
Would be nice to see if we can show this as a trend over time but will need adjusting for time of day and day of week
Data source as per note from Zella below
With regard to accessing the ED predictions, there are two tables on flow: