Flux for Data Visualization: A Preliminary Manifesto
summary
Argues for managing complex D3 visualizations with Flux, claiming explicit state beats buried D3 closures and event-driven DOM hacks. Describes a concrete implementation for the Tideline library used in Blip: D3 + Crossfilter for rendering and fast filters, a three-store architecture (config, data, draw) and dispatching interactions as actions so tooltips as actions replace DOM listener mutation. Shows working views (daily, trends, summaries), explains how stores coordinate load/filter/zoom, and labels the approach preliminary but practical.
AI-written summary · about
chapters (8)
- 0:00Flux for data visualization intro
- 3:57Sandwich metaphor for interfaces
- 8:31D3’s declarative vs procedural trade-offs
- 12:01D3’s getter-setter problems
- 15:53Why Flux fits complex visualizations
- 19:00Three-store structure for weather data
- 24:20Action flow in Flux pipeline
- 29:21Summary and future work
related talks
31mApplying Unidirectional Data Flow Beyond the ReactJS Ecosystem
59mReact, Flux and the Right Abstractions
25mA cartoon guide to the wilds of data handling in React
48mVisual Storytelling w/D3-A Conceptual Intro to Visualizing Data w/JavaScript
58mFrontin' like a Back-er
25mA Cartoon Guide to the Wilds of Data Handling
32mD3 and React, Together
51mD3 js: Easy and flexible data visualisation using Web standards