Weights

[tbeason]

It seems like there is no possibility for weights, but I know that other software packages (in R and Stata) allow for weighted quantile regression. I think it should not be too difficult, as the "unweighted" case is equivalent to equal weighting as you know.

Disclaimer: My use case involves weights because it is survey data.

[pkofod]

Yeah I think weights are quite easy to implement

[mcreel]

I have used kernel weights to do locally smoothed (e.g., nonparametric) quantile regression, relying on the interior point code from QuantileRegression.jl It is easy. See the following for a rough example
https://github.com/mcreel/ABCGMM/blob/master/DSGE/ABCGMM/npreg.jl

[pkofod]

2 years late to the party, but I've release the package now (put it in the registry) and I will add weights because I need them myself :) (actually I need nonparametric qr as @mcreel mentions, so I'll put that in as well).