Koutian Wu

Koutian Wu

AI4Sci PhD @ UT-Austin. ex-Undergrad @ USTC.

Geeta Persad Chronicle — SOTA Design Review

Geeta Persad Chronicle — SOTA Design Review

Goal: build a premium academic-chronicle page by borrowing proven patterns from open-source timeline, map-story, and editorial storytelling systems instead of inventing a bespoke interaction model too early.

1) Top 5 open-source references

1. TimelineJS3

  • Repo: https://github.com/NUKnightLab/TimelineJS3
  • Demo: https://timeline.knightlab.com/
  • Why it matters:
    • Best-known open-source editorial timeline pattern.
    • Strong slide-by-slide chronology with date precision, media support, and narrative pacing.
    • Good reminder that the timeline should be story-first, not just a dense data table.

2. StoryMapJS

  • Repo: https://github.com/NUKnightLab/StoryMapJS
  • Demo: https://storymap.knightlab.com/
  • Why it matters:
    • Canonical open-source “map + narrative card” reference.
    • Useful for the life-journey / geography layer: place, caption, image, transition.
    • Especially relevant because Sean’s biography has clear location chapters (China/Hong Kong -> Irvine -> Silicon Valley -> Canada / global ecosystem).

3. Odyssey.js

  • Repo: https://github.com/CartoDB/odyssey.js
  • Demo: https://cartodb.github.io/odyssey.js/
  • Why it matters:
    • Strong precedent for scroll-driven map storytelling.
    • More editorial than StoryMapJS: narrative blocks can drive map state, not just simple slides.
    • Helpful pattern for tying chronology and geography into a single reading flow.

4. vis-timeline

  • Repo: https://github.com/visjs/vis-timeline
  • Demo: https://visjs.github.io/vis-timeline/examples/timeline/
  • Why it matters:
    • Not as editorial out of the box, but excellent for dense professional timelines.
    • Strong zooming, grouped items, ranges, and interaction model.
    • Useful when the page needs a second, more “reference-grade” timeline view beyond the hero narrative.

5. Pageflow

  • Repo: https://github.com/codevise/pageflow
  • Why it matters:
    • Open-source multimedia storytelling system with a more museum / magazine sensibility.
    • Valuable less for direct adoption and more for structure: chaptering, full-bleed media, pacing, transitions, and premium editorial framing.
    • Good reminder that the site should feel like a curated exhibition, not a startup landing page with a timeline bolted on.

2) What we should directly adopt vs avoid

Directly adopt

  • From TimelineJS3:
    • One clear chronological spine.
    • Each milestone as a self-contained story card with headline, date, short narrative, and optional media.
    • Horizontal/sectional pacing principle: one important beat at a time.
  • From StoryMapJS:
    • A chaptered geography layer.
    • One location per life phase, with concise explanatory copy.
    • Camera transitions that move only when the story changes.
  • From Odyssey.js:
    • Scroll or chapter progression driving map/timeline state.
    • Sticky visual pane + narrative pane layout.
    • Treat map movement as narrative punctuation, not constant motion.
  • From vis-timeline:
    • Secondary “archive/reference” mode for power users.
    • Grouping milestones into phases: Education, Research, Founder, Ecosystem, AI Era.
    • Range bars for long eras, points for pivotal moments.
  • From Pageflow:
    • Full-bleed chapter breaks.
    • Strong rhythm: intro -> formation -> research -> company building -> ecosystem proof -> present/future.
    • High production-value feeling through spacing, restraint, and media hierarchy.

Avoid

  • Do not ship raw TimelineJS3 or StoryMapJS embeds as the main experience.
    • They are useful references, but the default styling feels educational / newsroom / template-like, not premium-founder-editorial.
  • Do not over-animate the map.
    • Constant camera movement makes biography pages feel gimmicky.
  • Do not make every partnership logo a milestone.
    • Too many enterprise badges turns the story into a press-release wall.
  • Do not force users into a single giant scrollytelling path with no quick navigation.
    • Founder pages also need a skimmable mode.
  • Do not use glossy “toy” sci-fi styling without evidence.
    • The current prototype has energy, but the final version should feel more archival, elegant, and source-backed.

3) Recommended information architecture for the Sean Xiang page

A. Hero / thesis

  • Name, one-sentence thesis, portrait or restrained abstract visual.
  • Short framing line: early technical prodigy -> research -> enterprise security founder -> AI infrastructure era.
  • 3-5 proof chips only, not 10+.

B. At-a-glance chronology

  • Compact horizontal phase rail:
    • Early formation
    • CUHK / PhD
    • Beckman / UC Irvine
    • Bloombase founding
    • Enterprise validation years
    • AI-era repositioning
  • Clicking a phase jumps to that chapter.

C. Main narrative chapters

Each chapter should have:

  • date / time range
  • chapter title
  • 120-220 words of narrative
  • 1-3 key proof points
  • optional image / logo / source link

Recommended chapters:

  1. Early acceleration
  2. Research and academic formation
  3. California transition
  4. Founding Bloombase
  5. Enterprise trust and ecosystem validation
  6. Cloud / platform expansion
  7. AI infrastructure chapter
  8. Present direction / why it matters now

D. Geography module

  • A separate but integrated map section.
  • 4-6 key place markers only.
  • Each place corresponds to a chapter, not every event.
  • Map should answer: how did the journey move across institutions and markets?

E. Validation / ecosystem section

  • Curated partner wall with context.
  • Better structure than a raw logo grid:
    • Cloud platforms
    • Security / HSM partners
    • Enterprise ecosystem validation
  • Keep it evidence-driven and selective.

F. Archive / sources

  • Expandable source list or evidence drawer.
  • This is what upgrades the page from tribute-style to museum/editorial credibility.

4) Visual / UX principles to make it feel premium, not toy

  • Favor editorial elegance over startup-dashboard gloss.
    • More off-white, charcoal, muted blue/green accents.
    • Less neon, fewer decorative orbit effects.
  • Build around typography first.
    • Premium biography pages feel expensive because the text hierarchy is calm and deliberate.
  • Use motion sparingly.
    • Fade, parallax, and camera transitions only at chapter boundaries.
  • Give milestones breathing room.
    • Fewer cards, more substance per card.
  • Distinguish “major chapter” vs “minor proof point.”
    • Not everything deserves equal visual weight.
  • Use maps as orientation, not spectacle.
    • Clean basemap, limited markers, subtle routes.
  • Prefer real artifacts over abstract decoration.
    • Photos, scans, conference visuals, article snippets, logos, citations.
  • Maintain dual reading modes:
    • skim mode for busy visitors
    • deep mode for readers who want the full chronology
  • Make source credibility visible.
    • Small citations and outbound references create trust fast.

5) Exact build recommendation: what stack to use now

Recommendation

Build this now as a lightweight custom site, not as an iframe composition of third-party tools.

Stack

  • Framework: Astro
  • UI islands: React (only where interaction is needed)
  • Styling: Tailwind CSS or plain CSS modules (either is fine; Tailwind is faster if the team wants speed)
  • Map: MapLibre GL JS
  • Scroll/chapter behavior: custom lightweight intersection-observer logic inspired by Odyssey.js / editorial scrollytelling patterns
  • Timeline rendering:
    • custom chapter timeline for the main narrative
    • optional vis-timeline secondary view for “full chronology / archive” mode
  • Content source: structured JSON or YAML in-repo for milestones, chapters, locations, logos, and sources
  • Deployment: static build on GitHub Pages / Vercel / Netlify

Why this stack now

  • Astro keeps the site mostly static, fast, and easy to ship.
  • React islands let us add only the interactive map/timeline pieces.
  • MapLibre avoids lock-in and is enough for a polished biography map.
  • A custom narrative timeline will look far better than raw TimelineJS styling.
  • vis-timeline can be added later only if a dense reference view becomes necessary.
  1. Define a clean content schema first:
    • chapters.json
    • locations.json
    • timeline-events.json
    • sources.json
  2. Build the page as 5-8 curated chapters, not dozens of equal cards.
  3. Sync chapter scroll state with the map.
  4. Add a compact phase rail for jumping.
  5. Add an archive drawer / sources panel last.

Bottom line

The right move is: borrow the information architecture of TimelineJS3 + StoryMapJS + Odyssey.js, but implement a custom editorial surface. That gives us proven storytelling structure without inheriting the dated visual language of off-the-shelf embeds.