NEWS.md

In carlbfrederick/ohiCovidMetrics: Used to calculate sub-state metrics relating to the COVID-19 pandemic

ohiCovidMetrics 0.8.9.2

Enhancements

• Updated population from WISH 2018 to NCHS 2020 estimates

ohiCovidMetrics 0.8.9.1

Bug fixes

• Fixed a bug in calculating total hospital beds and immediate beds available. I had been summing across columns and then across rows ... but this meant that if there was one NA in the row, the entire row was NA this silently dropping beds that we have been wanting to count. In technical terms we fixed sum(col1 + col2 + col3 + col4, na.rm = TRUE) into sum(col1, na.rm=TRUE) + sum(col2, na.rm=TRUE) + sum(col3, na.rm=TRUE) + sum(col4, na.rm=TRUE)

ohiCovidMetrics 0.8.9

Updates to column names and testing functions/calculations

Enhancements

• Made "Critical" label the default for CLI and Confirmed cases

Breaking Changes

• Removed the Testing Targets from the system per decision by the group since our source for some of these numbers is no longer being updated and there are no plans to include things like this in the dashboards.
• Removed the following deprecated functions:
  • pull_histTable_SG()
  • OLD_process_confirmed_cases()
  • clean_total_ed()
  • shape_total_ed_data()
  • process_total_ed()
  • rev_cusum_ucl()
  • rev_cusum_lcl()
  • calc_num_tests()

Bug fixes

• Changed API for historical data table again.
• Changed "Critical" to "Critically high"

ohiCovidMetrics 0.8.8

Updates to column names and testing functions/calculations

Enhancements

• Changed column output names in preparation for disseminating metric files via open data portal
• Changed testing percent positive calculation to be based on latest 7 day average percent positive and that this is the number the summary output is based on.
• Added a critical label to burden and composite classification functions in preparation for needing this but right now this ability is overwritten by 'Very high' by default.

ohiCovidMetrics 0.8.5.9

Bugfix to more robust way of truncating datetimes into dates.

Bug fixes

• For ILI and CLI metrics, switched from as.Date(POSIXt, tz = "America/Chicago") to as.Date(trunc(POSIXt, units = "days"))

ohiCovidMetrics 0.8.5

The major change in this update is on how we treat dates for confirmed cases. Based on stakeholder feedback, we now present daily confirmed case numbers with the same date that they are posted with in Historical Data Table and other displays.

Enhancements

• Changed Burden and Composite labels from 'Critical' to 'Very high' and made capitalization consistent across all labels.
• Daily confirmed cases will now match dates on the rest of the DHS reporting websites.
• Streamlined data check tests so that the bulk of the tests in the appended files are not duplicates.

Breaking Changes

• Merged files now contain 15 rows since we have 2 14-day windows that are off by a day. This caused many headaches. Also, summary rows are now dated as the maximum date in the data period to ensure there is only one summary row per geo-unit per data period.

Bug fixes

• Many, vastly improved tests for data and calculating functions

ohiCovidMetrics 0.7.9.9

This version corrects a critical bug in the calculation of the trajectory class.

Enhancements

• [none so far]

Breaking Changes

• [none so far]

Bug fixes

• Fixed bug in class_trajectory() that was expecting a ratio (old) instead of a percent change (current).
• Fixed issue in ILI where some regions were not merging thresholds and baselines correctly.
• Changed Testing Composite outcome labels

ohiCovidMetrics 0.7.9

This version is intended to be the last clean up version before we go live for the 'Tier 2' local dashboard.

Enhancements

• added some basic checks for merged metric file with basic reporting via a warning for failed tests.
• added function to aggregate new metrics onto source file. This function also re-calculates the moving average variables so that there are not breaks at the beginning of each update period.
• changed rounding for case trajectory/burden so that they don't look like true counts

Breaking Changes

• Have marked the Total Emergency Department visit functions as deprecated. These functions will be deleted before the 'Tier 2' local dashboard version of the file.

Bug fixes

• Cleaned up fix to end dates for testing sub-functions
• Fixed a couple of bugs in the ILI functions for calculating moving averages and merge fail for some HERC regions due to name variants.

ohiCovidMetrics 0.7.0

This large jump represents work such that all metrics topics are functional, but still need to confirm that the output is identical for 1.0.0 version

Enhancements

• added function to pull confirmed case metric directly from WEDSS
• fixed 'tests' in the Vignette introduced by output data format changes
• added functions to compute metrics for
  • Hospital Capacity
  • CLI (Covid Like Illnesses)
  • ILI (Influenze Like Illnesses)
  • Testing
  • Total Emergency Department Visits
• Added function to merge metric output together to streamline Tableau build
• All functions now output summary and daily metrics for Tableau dashboard
• Updated documentation so that all output variables of the process_*() functions are described.

Breaking Changes

• moved the clean_*() functions into the pull_*() functions
• moved the shape_*_data() functions into the process_*() functions to simplify user API
  • However, I have temporarily added the OLD_process_confirmed_cases() function to the package to ease transition to exclusive use of this package to produce the metrics.

Bug fixes

• Fixed a bug in rolling_weeks() in which periods that were a multiple of 7 days the earliest week was 7 NAs instead of a count. Now there can only be 0-6 NAs to start the week vector.
• Fixed bug in pull_histTable() where some fields were incorrectly imported as character instead of integer

ohiCovidMetrics 0.1.2

This patch fixes a small bug in the process_case_metrics() function that wrote over the trajectory scores with trajectory classifications.

Also:

• Fixes bug where some numeric columns were downloaded as factors in pull_histTable()
• Change license to GPL 3

ohiCovidMetrics 0.1.1

Minor patch to make output from process_case_metrics() more suitable for Tableau Ingestion.

ohiCovidMetrics 0.1.0

This is a minimally viable product that can download the Historical Data Table from the DHS website and produce the Confirmed Case Metric file for export to Tableau for the Dashboard. It was tested against the file that Jeff produced for the DHS dashboard for the period ending 17 June 2020 and matches exactly (see the Producing Confirmed Case Metrics Vignette).

The only data source supported is the Historical Data Table that is posted on the DHS website. The calculations are the ones described in the OHI Confirmed Case Metric Reproducibility document. Note: they are the ones described as 'defined as' not the ones that are 'proposed'.

ohiCovidMetrics 0.0.0.9000

• Added a NEWS.md file to track changes to the package.

carlbfrederick/ohiCovidMetrics documentation built on Jan. 10, 2022, 12:20 p.m.