Open davescroggs opened 2 years ago
For key stats explore if magnitude, proportion or percentile rank is more differential
roc_auc
numeric_gathered %>% group_by(metric) %>% roc_auc(rain_tomorrow, value, event_level = "second") %>% arrange(desc(.estimate)) %>% mutate(metric = fct_reorder(metric, .estimate))
https://github.com/dgrtwo/data-screencasts/blob/c914b888df88198b67e2898216f8c6d40c072e2a/ml-practice/ep4.Rmd#L130
For key stats explore if magnitude, proportion or percentile rank is more differential