Brainstorm and identify ways/metrics to analyze/compare manifests and manifests within elections

[MarcoStrijker]

This issue is a place to share possible ideas to analyze the manifests.

Current ideas:

Counts:

• [x] Number of pages
• [x] Number of sentences
• [x] Number of words
• [x] Number of characters
• [x] Number of syllables
• [x] Number of nouns
• [x] Number of adjectives
• [x] Number of verbs
• [x] Number of unique words
• [x] Number of hard words (words with more than two syllables)
• [ ] Number of unique hard words
• [x] Number of pronouns
• [x] Number of prepositions (https://en.wikipedia.org/wiki/Adpositional_phrase)
• [ ] Number of postpositions (https://en.wikipedia.org/wiki/Adpositional_phrase)
• [ ] Number of circumpositions (https://en.wikipedia.org/wiki/Adpositional_phrase)

Averages:

• [x] Average number of sentences per page
• [x] Average number of words per sentence
• [x] Average number of characters per sentence
• [x] Average syllables per sentence
• [x] Average syllables per word
• [x] Average nouns per sentence
• [x] Average adjectives per sentence
• [x] Average tree depth per sentence (https://en.wikipedia.org/wiki/Parse_tree)

Indexes/Other:

• [ ] Adjusted Flesch Douma index
• [ ] Gunning Fog index (https://en.wikipedia.org/wiki/Gunning_fog_index)
• [ ] Entropy (https://www.princeton.edu/~wbialek/rome/refs/shannon_51.pdf)

Vector:

• [ ] Average distance between sentences within program
• [ ] Average distance beteeen sentences between programs

[MarcoStrijker]

The number of capital letters also would be an interesting metric

[MarcoStrijker]

Garden-path sentences would be also interesting (https://en.m.wikipedia.org/wiki/Garden-path_sentence)