A year of reading, 2025

Did I actually read this year, when, and on what?

The pipeline

Question	One sentence, plain language	written before any chart
Data contract	notebooks/01, cell 2	variables, types, units — decided before chart type
Evidence	raw CSV → DuckDB/Polars → processed CSVs	aggregations land in files, not memory
Chart selection	notebooks/01, audit cell	3 candidates per chart, rejection reasons recorded
Charts	charts/*.html via lieflat-en	F2 hairline line · F1 rung bars · F4 tick donut — one silhouette each
Essay	posts/2025-reading-year-draft.md	the deliverable the charts serve

The proof

The full write-up lives in the essay; here is the chart that started it — October as the high-water mark, with the July dip shown rather than cropped.

Artifacts

• notebooks/01-template-tour.ipynb — data contract, DuckDB + Polars aggregates, chart-selection audit
• data/raw/reading-log-2025.csv — 291 sessions, deterministic (seed 20250101)
• data/processed/*.csv — monthly minutes, genre minutes, genre share
• charts/*.html — three single-file charts, source lines naming type, unit, dataset, notebook
• posts/2025-reading-year-draft.md — the essay draft
• run_notebook.py — one-command rerun of the whole pipeline

This project is the worked example of the project template this site's pipeline is built on: same folders, same conventions, same chart engine.