Stylography
Stylography is the study of writing patterns—the fingerprints left behind in prose. It’s not about what you say, but how you say it. This analysis examines sentence structure, voice, emotional arc, and the words that appear again and again across a corpus of work. Think of it as a data-driven mirror: the rhythms, the choices, the obsessions. Here’s what the numbers reveal about how I write.
Statistics
Word counting is trickier than you’d think. Microsoft Word counts hyphenated words as one. Google Docs splits them into two. Scrivener has its own opinions. Medium rounds up for reading time but won’t tell you the actual count. Even the Unix wc command differs from Python’s tokenizers. For this analysis, word counts come from spaCy’s tokenizer, treating contractions as single words (“don’t” = 1 word, not 2) and ignoring punctuation. It’s not perfect, but it’s consistent—and in stylography, consistency matters more than convention.
- Total Word Count: 219,027
- Number of Articles: 249
- Average Article Length: 880 words
- Longest Article: Zettelkasten Wasn’t Built in a Day (3,365 words)
- Shortest Work: A Princess (89 words)
How I Write
Readability scores come from the Flesch formulas, which measure how easy text is to read based on sentence length and syllable count. Reading Ease ranges from 0 (very difficult) to 100 (very easy)—higher is simpler. Grade Level estimates the U.S. school grade needed to understand the text. A grade 6 score means a sixth grader could follow along. These aren’t perfect (they ignore context, vocabulary complexity, and narrative structure), but they’re useful benchmarks for tracking consistency across a body of work.
- Average Sentence Length: 11.58
- Reading Level: 6.16
- Reading Ease: 71.52
How I Talk To You
Voice metrics reveal rhetorical patterns. Being direct (3.73 imperatives per 1,000 words) is moderate—occasionally direct without being pushy. Questions (6.04 per 1,000 words) show moderate use of questions to engage readers. The authority/vulnerability split indicates how often confident assertions appear versus admissions of uncertainty. Opening and ending patterns show no dominant formula—each piece finds its own shape rather than following a template.
- Being Direct: 3.73 per 1,000 words
- Questions I Force You to Answer: 6.04 per 1,000 words
- Voice Stance: ~67.74% authority / ~32.26% vulnerability
- How I Open Pieces: No Pattern (239, 96%), Question Hook (7, 3%), Statistic/Claim (2, 1%), Personal Scene (1, 0%)
- How I End Pieces: No Pattern (247, 99%), Rally (1, 0%), Warning (1, 0%)
Emotional Shape
Emotional shape tracks sentiment and volatility. Average tone (0.018) stays near neutral—balanced and measured, neither overly dark nor sunny. Volatility (0.119) shows some emotional variation within pieces. “Steady” endings maintain the piece’s emotional register rather than building to a crescendo or crashing to despair. It’s consistency, not monotony. Most pieces stay emotionally steady rather than manipulating the reader toward a specific feeling at the close.
- Average Emotional Tone: 0.018 (0 = neutral, negative = darker, positive = hopeful)
- Emotional Volatility: 0.119
- How I Tend to Land: Steady (232, 93%), Falling (9, 4%), Rising (8, 3%)
Receipts
The receipts index measures how often writing explicitly names companies, people, or systems rather than speaking in generalities. It tracks proper nouns and entities that could be fact-checked or challenged. Scores below 0.5 are low (rarely specific), 0.5-1.5 are moderate (balanced use of examples), and above 1.5 are high (frequent call-outs). A score of 0.990 is moderate—occasionally references specific companies, people, or systems. Higher scores suggest writing that’s grounded in specific examples rather than abstract arguments. It’s the difference between “tech companies exploit users” and “Facebook sold user data to Cambridge Analytica.”
- Receipts Index: 0.99
What I Can’t Stop Talking About
Lexical diversity metrics reveal overall vocabulary habits across the corpus. Type-token ratio (0.121) is low—writing favors a consistent core vocabulary. This isn’t good or bad—Hemingway had a famously limited vocabulary. It’s about clarity versus ornamental variety. Hapax ratio (0.42) counts words used only once. Together, these show whether writing leans on familiar foundations or reaches for constant novelty. The words that appear most frequently across the corpus:
- time (in 210 posts)
- year (in 188 posts)
- think (in 184 posts)
- work (in 181 posts)
- good (in 180 posts)
- come (in 180 posts)
- know (in 172 posts)
- day (in 171 posts)
- find (in 168 posts)
Obscurity
- fisherman → the-fisherman
- synergy → synergy
- centralization → centralized-decentralization
- sebastian → la-la-land
- leach → charlesdickensredraiders
- talkie → walkie-talkie-and-netflix-originals
- walkie → walkie-talkie-and-netflix-originals
- unshakeable → unshakeable-financial-freedom
- expectancy → risk-and-time
Lexical Diversity: 12035 unique base words from 99583 total tokens. Type-token ratio: 0.121. Hapax ratio: 0.417.
Article Recommendations
Each article includes algorithmically-generated recommendations for related reading—a mathematical approach to discovering connections across the corpus. The recommendation engine analyzes content similarity and suggests pieces that complement or extend the current topic. This creates pathways through the writing that readers might not find on their own.
Cornerstone Article: Presenting The Day Life Breaks, A Crowdsourced Affair appears as the top recommendation 5 times across 247 posts (2.0%), making it the central connecting point in the corpus.