Menu
▾
▴
#474 smooth unique standard deviation
Hey!
I'm using smooth unique to plot the average y data points with same x values. I was wondering if there is an easy way to not only get the average but also the standard deviation to easily draw some errorbars?
I know smooth is probably not the ideal way, but since I dont see any other way (without further processing the data), a pseudocolumn output of the StdDev would be great :)
What do you think?
Cheers,
Alex
If your data is sparse (only a few x values have multiple representatives) then I do not think such a plot is a good representation - better to just plot all the points.
If your data is dense (many y values for each x value) then I suggest using boxplots. In the limiting case you could use one boxplot for each x value. For somewhat less dense data you could maybe bin the data first and then use one boxplot per bin.
Hey Ethan,
thanks for your answer.
My data is more like your first example, so sometimes I plot all the points and use the smooth option to draw a line to guide the eye. But sometimes I rather plot only the average values with errorbars and I thought this might be interesting for more people.