It has come time to define the final scope of the project and set a roadmap before the project is concluded.
Data reporting ended December 31, 2023. Some datasets received retroactive data updates as late as April 12, 2024, but no new data were reported after the end of 2023.
Roadmap
Existing datasets
Maintenance of existing datasets (end date: TBD)
Audit of historical datasets (#118)
Add alternate PT-level datasets for cases/deaths where they offer superior accuracy/temporal resolution compared to aggregated HR-level data (#120)
Update names of hospitalization and ICU data (hosp_occupancy and icu_occupancy) (#132)
Attempt to update NB death time series to account for correction on 2022-10-18; verify if PHAC dataset has corrected bump on 2022-12-17 (originally discussed in #80) (#118)
Roadmap: Rejected ideas
Some ideas that have been proposed over time are unfortunately beyond the scope of what I am able to do with the time and resources I have. If time/money/personnel became available in the future, these ideas could be reconsidered, but for now these issues will be closed.
Additional processing of data
68
This would be a useful addition, but it would take a tremendous amount of work that would likely require individual attention to each input dataset. It would be feasible to automate this for archive_ts-type datasets, but many of these would have to be redone since they have been preprocessed into filled time series using helper_ts.
1
Additional postprocessing over raw data to remove impossible values would be useful, but creates its own problems. Better to document the existence of impossible values and explain some reasons for them. (#93)
Extending existing datasets
69
Collectively, Horizon Health Network and Vitalité Health Network should represent the total hospital and ICU occupancy values for COVID-19 in NB. The province stopped reporting this value in December 2022, whereas the two health networks continue to report it. However, at the end of the provincial time series, hospital occupancy numbers seemingly included only "for" COVID, whereas the Health Network values have always included "with" and "for". In addition to being a lot of effort to collect historical values, it would be jarring to switch back to the more expansive definition.
Additional sub-HR datasets
13
ccodwg/Covid19CanadaDataProcess#42
ccodwg/Covid19CanadaETL#42
Other additional datasets
Vaccination in last 6 months (#91)
This metric was reported for less than a year (2022-08-14 to 2023-06-18) and so is de-prioritized
44
45
ccodwg/Covid19CanadaETL#48
This would be useful, but a significant undertaking for limited gain. The exception could be Quebec (#79), but it would be much simpler to simply ensure that the QC dataset drops out gracefully at the appropriate time.
19
Alternative sources of data other than mainline, government-reported metrics would be interesting, but beyond the scope of this project
Variant data (#70)
I could add a Canada-level variant time series from archived versions of the PHAC dataset (909a7c17-2773-4536-add4-717b59deea4c), although it has changed format at least once. The "other" category may also pose an issue, particularly if it gets redefined over time (or retroactively).
It has come time to define the final scope of the project and set a roadmap before the project is concluded.
Data reporting ended December 31, 2023. Some datasets received retroactive data updates as late as April 12, 2024, but no new data were reported after the end of 2023.
Roadmap
Existing datasets
Additional datasets
Additional processing of data
Documentation
93
Other: Long term
hosp_occupancyandicu_occupancy) (#132)Roadmap: Rejected ideas
Some ideas that have been proposed over time are unfortunately beyond the scope of what I am able to do with the time and resources I have. If time/money/personnel became available in the future, these ideas could be reconsidered, but for now these issues will be closed.
Additional processing of data
68
archive_ts-type datasets, but many of these would have to be redone since they have been preprocessed into filled time series usinghelper_ts.1
Extending existing datasets
69
Additional sub-HR datasets
13
Other additional datasets
44
45
19
909a7c17-2773-4536-add4-717b59deea4c), although it has changed format at least once. The "other" category may also pose an issue, particularly if it gets redefined over time (or retroactively).