gistfile1.pyw

#!/usr/bin/env python
from __future__ import division
import nltk

nltk.data.path.append('/Users/mike/.nltk_data')

inaugural = nltk.corpus.inaugural

with open('./sentence_lengths.csv', 'w') as out:
    for fileid in inaugural.fileids():
        avg_len = (len([w for w in inaugural.words(fileid) if w.isalpha()]) /
                   len(inaugural.sents(fileid)))
        out.write("%s,%s\n" % (fileid[0:4], avg_len))