word count frequency except punctuation

Question

I want to count the 20 words that appear most frequently in the dataset, except for punctuation. So far the code I have tried is

dataset[\'token\'].value_coun