These unigram frequencies are taken from Google Web Trillion Word Corpus.

I found them hosted at https://norvig.com/ngrams/count_1w.txt

I took the 100,000 most frequent words/unigrams from this file, and two are randomly selected. Obscurity level changes which first k words are sampled. High: 100k, Medium: 20k, Low: 4k.

Check out a version I made here for Chinese Hanzi.