Changelog

mqor 0.1.0

This is a minor version update to signal that the development of mqor is entering into a new phase of active use throughout the validation project.

New Features

• The default_params have been once again updated based on outcomes from Task 3 of the European Standard on modelling quality objectives for assessment of ambient air (EC-ENV/2024/OP/0036).

Bug Fixes

• Updated mutate_rolling_mean() to use has.nf over hasNA, which has been deprecated.

mqor 0.0.17

New Features

• validate_coverage() now returns padded data, which will help for future use in, e.g., mutate_rolling_mean().

• pad_data() is now exported, which pads data without performing the coverage validation.

• Automatic assumptions of time resolution (e.g., in validate_coverage()) are now much quicker.

mqor 0.0.16

Bug Fixes

• Reinstated the min_coverage argument for summarise_daily() and summarise_annual().

• Fixed an error which caused mutate_rolling_mean() to be overly strict with missing values.

mqor 0.0.15

New Features

• Added validate_coverage() which ensures short-term data meets data capture thresholds.

  • Removed the min_coverage argument from summarise_annual() and summarise_daily(); this is now dealt with by the validate_coverage() argument.

• Added ‘resolution’ and data coverage % options to the Shiny interface. This allows users to define whether their short-term data is hourly or daily, which will be used to filter both the short- and long-term data.

  • The resolution can also be “other”. This disables data coverage filtering in the mqor interface.

  • ‘Resolution’ purely reflects the time sequence of the data. If a user inputs 8-hour running ozone, this is still hourly data. Similarly, maximum daily 8-hour running ozone is daily data.

  • If a user provides hourly data with an intention to calculate a daily average within the mqor Shiny interface, they should still indicate that the data is hourly; data capture thresholds are applied on the ‘raw’ data before any averaging is done.

  • Temporal filtering of short-term data (e.g., for certain seasons) is likely to cause issues with data capture calculations. If a user is filtering for certain time periods, consider setting resolution to ‘other’ to disable data capture thresholds entirely.

• Uploading ‘MQOR format’ data to the Shiny interface now allows users to specify a different delimiter. This still defaults to ";".

• The Shiny interface now fails more gracefully if the user filters away all of their data.

• All plotting functions gain the title argument, which adds an optional title to the plot.

mqor 0.0.14

New Features

• plot_mqi_scatter() gains censor for static short-term radar charts.

• plot_mqi_scatter() now annotates the four quadrants in short-term radar charts.

• summarise_mqo_stats() now includes n (the number of sites) in the summary object, and will now warn if the user provides fewer than 10 sites.

  • The Shiny interface now also warns if the user provides under 10 sites.+

mqor 0.0.13

Bug Fixes

• plot_mqi_report() now more consistently uses absolute values when plotting 90th percentiles and applying censor.

mqor 0.0.12

New Features

• If tasked with assuming the MQI parameters to use, summarise_mqo_stats() will now print these as a message to alert a user which parameters have been selected.

• show_annotations now includes the pollutant and year.

Bug Fixes

• plot_mqi_bars(show_annotations = TRUE) will now place legends and annotations in appropriate places.

mqor 0.0.11

New Features

• plot_mqi_report() has gained show_annotations.

• plot_mqi_bars() now has a line type legend (for static plots) and a more descriptive plot annotation (for interactive plots) to better describe the horizontal annotations being shown.

• summarise_mqo_stats() now validates that only a single year of data has been provided, and returns the input year of data in its params.

• Various plotting functions now verify that stats and stats2 have the same year and pollutant.

• The fontsize argument within gp now controls the size of annotations when interactive = TRUE.

Bug Fixes

• Fixed an issue determining the resolution of input data for automatic parameter identification.

• plot_comparison_bars() now uses RMSU0√(1+β²) as legend labels rather than RMSU0*, reflecting changes in the CEN specification.

• Annotation size has been reduced to better fit annotations on most plots.

• plot_mqi_bars() and plot_comparison_bars() now check for x-axis label overlaps and will remove overlapping axis labels.

mqor 0.0.10

Bug Fixes

• Fixed an issue preventing the Shiny dashboard being launched.

• Fixed a bug in mutate_rolling_mean() which would underestimate data capture.

mqor 0.0.9

New Features

• The default_params have been updated based on outcomes from Task 3 of the European Standard on modelling quality objectives for assessment of ambient air (EC-ENV/2024/OP/0036).

• The Shiny GUI can now use the demo_shortterm and demo_longterm datasets by selecting “DEMO” on the first screen.

• The Shiny GUI now shows the version of MQOR in the navigation bar. This links to the MQOR change log.

• When show_annotations = TRUE, the MQOR version is also included as an annotation.

mqor 0.0.8

New Features

• [BREAKING] stats_shortterm and stats_longterm have been renamed stats and stats2 wherever they appear. Long- and short-term data will be automatically determined by each function and used appropriately. While this may break existing code, future code will be more streamlined for the user who can simply provide short- and long-term statistics in any order.

• Most plots have gained the show_annotations argument, which allows for the input parameters and complementary indicators to be appended to a greater number of MQI evaluation plots, and for it to be disabled in plot_mqi_scatter().

mqor 0.0.7

New Features

• Filtering options have moved around to be in a more sensible order.

• The existing temporal filtering options are now exclusive to short-term data.

• A user can now select whether they want to use their uploaded long-term data or have the app time-average their short-term data.

• If the user selects to time-average the short-term data, they can additionally apply an independent month filter on the short-term data. This allows for seasonal averages to be calculated for pollutants like ozone.

mqor 0.0.6

Bug Fixes

• plot_mqi_scatter(interactive = TRUE) will now correctly label the U_RV value.

mqor 0.0.5

Bug Fixes

Fixes from resolving disparities in the technical specification:

• vec_ti_cor() is now calculated using the correct formula.

• plot_mqi_scatter() will now put TI_CRMSE on the correct segment of the plot (the inverse of mqor 0.0.4).

mqor 0.0.4

Bug Fixes

• vec_ti_crmse() is now calculated using vec_crmse() and vec_rmsu() instead of vec_mqi() and vec_pi_bias().

• plot_mqi_scatter() will now put TI_CRMSE on the correct segment of the plot.

• plot_mqi_scatter() will now extend the extents of the target diagram if either of |TI_CRMSE| or |TI_BIAS| are greater than 2.

mqor 0.0.3

New Features

• default_params has been updated based on recent revisions of the technical document.

• Web documentation has been significantly expanded.

Bug Fixes

• The units of reference values in default_params now vary, and the dataset now has a rv_unit column to correctly label them.

mqor 0.0.2

Bug Fixes

• read_delta_resource() will now ignore the case of startup.ini column headers.

mqor 0.0.1

• Development version as part of the European Standard on modelling quality objectives for assessment of ambient air (EC-ENV/2024/OP/0036, European Commission)