Building a Grammar for Statistical Graphics in Clojure
summary
A grammar-of-graphics implementation in Clojure and ClojureScript that treats plots as data: specs are plain Clojure maps/records mapping variables to aesthetics, compiled into a scene graph compiled to SVG/PDF/PNG and tunable before rendering. Uses core.logic rule rewrites to expand concise specs (e.g., boxplot -> quantile stats, histogram -> binning + sum) so exploratory changes (swap x from weight to cylinders, switch points to boxplots, facet by group) are one-line edits rather than big rewrites. Examples: wind-farm iPad dashboard, gene-seq browser; positioned between Excel and D3.
AI-written summary · about
chapters (13)
- 0:00Start
- 2:18Wind energy
- 4:24Bioinformatics
- 8:45Some aesthetics are better than others
- 11:27Off the rack
- 12:39D3: Data Driven Documents (2011) Mike Bostock Jeffrey Heer Vadim Ogievetsky
- 13:13Bespoke
- 14:40The Grammar of Graphics (1999)
- 15:48Motor Trend cars dataset
- 18:51complected
- 25:27Pattern rewrites
- 27:44Consider the histogram
- 29:14API Support
related talks
39mCrafting Artisanal Vector Graphics
19mGrammar of Graphics in Python with Plotnine
15mViz clj - Clojure data visualization for beginners
13mggsql: A grammar of graphics for SQL
31mUsing rule engines for charts and poetry
31mLightning Talks
38mScientific Clojure, a bird's eye view
01h 02mSolving Problems the Clojure Way