we have an R package which you can use to access the data: https://github.com/langcog/wordbankr

we've done some analyses predicting words' learnability from frequency and other factors: http://langcog.stanford.edu/papers_new/braginsky-2016-underr...