05-04: Consider using mode instead of median for scaling flats?

[ejmcgrath]

I've always used the mode to scale (sky/twilight) flats since I worry about the median being biased high by stars. This is obviously more processing intensive, though, so there's a trade-off. Can be implemented with statistics.mode or scipy.stats.mode.

[mwcraig]

Good suggestions, I'll a note about using mode. One question: are you typically working with integer or float data when you use mode?

[ejmcgrath]

My raw data are always integer counts. I'm not sure how the statistics module (or the scipy module) deal with it if it's float. Good question.

[mwcraig]

Still a good idea and still needs to be added...hopefully soon 😬