Skip to the content.

presidential-readability — Dataset Codebook

Generated: 2026-11-15

Columns

president

speech_date

year

title

speech_type

is_sotu_series

delivery_mode

url

source_file

fk_grade

flesch_reading_ease

smog_index

gunning_fog

coleman_liau

word_count

sentence_count

Notes

Source: University of Virginia Miller Center presidential speech corpus (public domain), https://data.millercenter.org/miller_center_speeches.tgz. Readability computed with textstat==0.7.13 (Flesch-Kincaid + companions). CAVEAT: readability formulas were built for written prose; scores on transcribed spoken speech are a consistent index, not a literal grade. Segment trends at the written->spoken (~1913) delivery break. Coverage density varies sharply by era.