[Bug Report]: plot_productivity_anomaly filling in missing years

[jcaracappa1]

Describe the bug

For plot_productivity_anomaly(report = 'NewEngland', plottype = 'region', EPU = 'GB') there is correctly a gap in the data due to missing surveys but when plot_productivity_anomaly(report = 'NewEngland', plottype = 'council', EPU = 'GB') those gaps appear to be filled in

Describe the situation

plots drawing from different data when plottype = 'council' is supposed to just be a different mix

Additional information

No response

[MGrezlik]

@jcaracappa1

Reason for bug
anomaly plots for plottype 'region' come from dat_spec_rec_epu_forSOE which comes from EDAB_Dev/beet/surveyBiologicalByEPUData.rds.

anomaly plots for plottype 'council' come from dat_spec_rec_forSOE which comes from EDAB_Dev/beet/surveyBiologicalData.rds

The raw survey data had observations in 2019 with both SPRING and FALL present as well as all EPUs. In 2020 the raw survey data had observations only in the SPRING in MAB.

In the workflow we filter out values for 2020 in dat_spec_rec_epu_forSOE but not in dat_spec_rec_forSOE here. That was done to match what was done in ecodata::get_productivity_anomaly but there was no explanation other than 'Select and rename'. Filtering out 2020 creates NA values for 2019 as we need small fish the following year to calculate an anomaly.

Fix options

• Filter out 2020 in dat_spec_rec_forSOE in the workflow
  • probably 'best practice' for workflow but would require Brandon replacing productivity_anomaly.rds in ecodata pre-production branch
• Filter out 2019 and 2020 in ecodata::plot_productivity_anomaly
  • kind of a patch but would not require generating a new productivity_anomaly.rds

[jcaracappa1]

Option A sounds more in line with my understanding of the concern of using 2020 survey data.

[MGrezlik]

@jcaracappa1
I made that change to productivity_anomaly in the workflows repo, generated a new productivity_anomaly.rds file (in EDAB_Dev/grezlik), then generated the plots. Since the 2020 data is filtered out after the anomaly is calculated, any species that was captured in that 2020 Spring MAB survey has an anomaly value for 2019 while all others do not. Should I filter out 2019 anomaly data as well to be thorough in avoiding any inclusion of 2020 data?

Updated Bug Plots
plot_productivity_anomaly(report = 'NewEngland', plottype = 'region', EPU = 'GB')

plot_productivity_anomaly(report = 'NewEngland', plottype = 'council', EPU = 'GB')

All Updated Plots
plot_productivity_anomaly(report = 'MidAtlantic', plottype = 'region', varName = 'anomaly', EPU = 'MAB')

plot_productivity_anomaly(report = 'MidAtlantic', plottype = 'council', varName = 'anomaly', EPU = 'All’')

plot_productivity_anomaly(report = 'MidAtlantic', plottype = 'region', varName = 'assessment')

plot_productivity_anomaly(report = 'MidAtlantic', plottype = 'council', varName = 'assessment')

plot_productivity_anomaly(report = 'NewEngland', plottype = 'region', varName = 'anomaly', EPU = 'GOM')

plot_productivity_anomaly(report = 'NewEngland', plottype = 'region', varName = 'assessment')

plot_productivity_anomaly(report = 'NewEngland', plottype = 'council', varName = 'assessment')

[jcaracappa1]

i don' think you need to filter out the 2019 MAB since the Spring 2020 was completed