Started this script to recreate the rankings and ran into some detail questions:
[x] using the rank() function in R recreates the rankings shown on the website stoplight table, but the implied direction is that lower values = better (i.e., higher rank). Does this apply the same way across all indicators? Aren't higher values "better for salmon" for some of the indicators? This is linked to issue #69 and the results presented in the Chasco et al. 2021 paper. For example, the copepod richness anomalies, are they calculated such that lowest number (-3.769 for 2000) corresponds to highest richness? Currently year 2000 is top-ranked on this indicator. Yes, the ranking is adjusted for each indicator based on assumed directionality. For 5 of them larger is better (DeepSalinityAvgMayToSept, Northern Copepod biomass, Nearshore Ichthyoplankton Jan-Mar, Chinook Juv biomass, Coho Juv biomass).added a table to the readme for this dataset.
[x] how to handle ties? (especially in mean ranks?) the ONI variable has a tie for 1999 and 2000, and the rankings on the website correspond to rank(x,ties.method = "min"), so using that.
[ ] how to handle missing values? The biological transition variable is missing several years, and the rankings on the website correspond to rank(x,ties.method = "min",na.last = TRUE), so using that.
[ ] Any early idea yet how the 2022 ocean entry Chinook and Coho stocks did, given that the climate indicators gave a very different picture than the temp and salinity indicators? Plans to revisit the mix of indicators and relative weighting once returns from those cohorts are fully assessed?
Started this script to recreate the rankings and ran into some detail questions:
rank()function in R recreates the rankings shown on the website stoplight table, but the implied direction is that lower values = better (i.e., higher rank). Does this apply the same way across all indicators? Aren't higher values "better for salmon" for some of the indicators? This is linked to issue #69 and the results presented in the Chasco et al. 2021 paper. For example, the copepod richness anomalies, are they calculated such that lowest number (-3.769 for 2000) corresponds to highest richness? Currently year 2000 is top-ranked on this indicator. Yes, the ranking is adjusted for each indicator based on assumed directionality. For 5 of them larger is better (DeepSalinityAvgMayToSept, Northern Copepod biomass, Nearshore Ichthyoplankton Jan-Mar, Chinook Juv biomass, Coho Juv biomass). added a table to the readme for this dataset.rank(x,ties.method = "min"), so using that.rank(x,ties.method = "min",na.last = TRUE), so using that.