The histogram is a really efficient and useful way to visualize and describe data. Lately, I have been wondering if there is an advantage to bootstrapping a histogram? i.e. instead of generating a single histogram, generate 1000 by resampling the data with replacement, then average them all (perhaps after a small amount of smoothing) to get a bootstrapped estimate of the data density.
I have not really found anyone discussing this, but I think this is because when I search for 'bootstrapped histogram' I mainly get results about histograms showing the results of a bootstrap on some other dataset.
Can anyone point me to a resource discussing this? Or tell me why it is not a good idea?