Key Takeaways: Theming and Branding

  1. Style is a system, not a one-off. The chapter's threshold concept: a visualization brand is a codified system (colors, fonts, templates, helper modules) that makes consistency automatic rather than effortful. Build once, apply everywhere.

  2. A style guide documents the rules. Colors, fonts, chart anatomy, encoding conventions, annotations, layout, accessibility, examples, do's/don'ts. The document serves as reference, training material, and specification for implementation.

  3. A color system has multiple palettes. Primary/secondary/accent colors, plus categorical (for groups), sequential (for ordered values), diverging (for values with a meaningful midpoint), neutral (for backgrounds and grids). Different data types need different palettes.

  4. matplotlib style sheets are .mplstyle files. Write rcParams values in a text file, install to matplotlib's config directory, apply with plt.style.use("name"). Compose with base styles: plt.style.use(["seaborn-paper", "mybrand"]).

  5. Plotly templates register in pio.templates. Create a go.layout.Template with your settings, register with pio.templates["mybrand"] = template, set as default with pio.templates.default = "plotly_white+mybrand". The + operator composes templates.

  6. Helper modules encapsulate conventions. A Python module with apply_brand(), branded_axes(), add_title(), add_source(), save_branded() functions. Authors invoke the helpers rather than remembering the rules.

  7. Adapt corporate brand guides. Extract colors, test colorblind safety, identify fonts, fill in chart-specific rules, build implementation, get design approval. The chart brand extends the corporate brand into a new medium.

  8. Compliance can be automated. Write validate_brand_compliance(fig) that checks font sizes, colors, backgrounds, and other rules. Run before publishing to catch drift. Optionally integrate into CI/CD.

  9. Packaging beats design. The BBC's bbplot succeeded because it packaged the style as a reusable library, not because the style itself was novel. Code beats prose for enforcing conventions. For your own brand, encode the rules as software.

  10. Brands evolve. Version the style sheet. Provide migration paths when updating. Document changes. Solicit feedback. A brand system is infrastructure that must be maintained, not a monument that stays static.

With Chapter 32 complete, the technical and systematic foundations of visualization production are in place. Chapter 33 integrates everything in the book into an 8-step workflow from analytical question to published chart, providing a process framework that ties together perception (Part I), design (Part II), code (Parts III-V), specialized techniques (Part VI), and production (Part VII).