Code review for public release.
-I have reviewed the code in the GitLab repository “code.usgs.gov/werc/dixon-field-station/Halstead-lab/giant-gartersnake-survival-analysis” and found no major issues. The code functioned as intended and produced results in the form of model comparison using AICc and two figures showing survival curves. Below are a few minor comments with suggestions for improving the clarity of the code.
-The code calls for the two csv files to be loaded from a folder named data but the repository does not contain such a folder. I recommend removing “data/” from the directory used to load the csv files. Alternatively, you could add a “data” folder to the GitLab repository, but because the data will not be stored in GitLab, I think it is better to remove “data/” from the directory.
-On line 82, instead of manually writing the AICc values for each model, it could be cleaner to call the AICc function as above on lines 71–79. For example:
aics <- c(AICc(ggs_cox_null), AICc(ggs_TG), AICc(ggs_T), AICc(ggs_act), AICc(ggs_TG_act), AICc(ggs_T_act),AICc(ggs_TG_dist), AICc(ggs_T_dist),AICc(ggs_dist))
-Correct “trtment” to “treatment” on line 34.
-Correct “transloacted” to “translocated” on line 36 (second use of word in the line).
-There is a typo on lines 86 and 170, should read “visualize” instead of “visulaize”
-The code ran without errors and produced the expected figures.
-The top model for adult survival according to AICc was “ggs_T” and included an effect of translocation on survival.
-The top model for juvenile survival according to AICc was the null model.
-It could be nice to have a few lines of code to produce an AIC model comparison table for each analysis. The data frame could have columns for model name, AICc, delta_AICc, rel_LL, and akaike_weight. I find that helpful for comparing models. Here is an example I put together for the adult survival analysis:
model.df <- as.data.frame(matrix(NA, nrow=9, ncol = 5))
colnames(model.df) <- c('model','AICc', 'delta_AICc','rel_LL','akaike_weights')
model.df$model <- c('cox_null','TG','T','T_act','TG_act','TG_dist', 'T_dist','act','dist')
model.df$AICc <- aics
model.df$deltaAICc <- akaike.weights(aics)$deltaAIC
model.df$rel.LL <- akaike.weights(aics)$rel.LL
model.df$akaike_weights <- akaike.weights(aics)$weights