Open kaitlyngaynor opened 4 years ago
Asked some folks about this at the workshop - there isn't an easy way to have separate scales, but it's also a really bad idea to do this because it can lead to spurious correlations appearing when there are none. Think about other ways to visualize.
Perhaps - proportion of all of the detections that occur in that month? Scale to average RAI? Log scale?
https://github.com/kaitlyngaynor/gorongosa-camera-traps/blob/d35c695e9e7e3fb172da392db907b1171a3abf1d/shiny-rai/server.R#L209-L215
Would be nice to have two separate scales for the y-axis for subset A and B.
Also, is it possible to keep width of bar consistent even when the value of one bar is 0? Currently, if there is only one bar for a given month, it is twice as wide.