šŸŽÆ BC AI Survey: Semantic Visualizations

Interactive 3D clustering of 1,001 British Columbian voices on artificial intelligence

Welcome to the BC AI Survey Data Storytelling Project!

This interactive dashboard presents semantic visualizations of open-ended responses from 1,001 British Columbians about artificial intelligence. Each 3D scatter plot represents survey responses as points in semantic space, where similar responses cluster together naturally.

šŸ” How to explore: Click on any visualization below to dive into the 3D interactive plots. Rotate, zoom, and hover over points to discover the authentic voices and perspectives of BC residents.

šŸŽ§ New in this version: Every question page now also includes an AI‑generated roundtable dialogue between cluster "representatives" (synthetic voices derived from thematic groupings of responses) followed by an automatically generated synthesis summary.

šŸ—ŗļø Semantic Maps Core Feature 1

3D t‑SNE landscapes of response embeddings let you visually browse the conceptual terrain of what people said.

  • šŸ“Œ Each point = a unique textual response (aggregated by identical text, sized by frequency).
  • šŸŽØ Colors = data‑driven clusters (K‑Means, silhouette-optimized) of semantically similar ideas.
  • 🧭 Hover for raw wording + cluster label; rotate & zoom to see structural patterns.
  • šŸ“ˆ Size uses cube‑root scaling so very frequent responses stand out without dominating.

Open any question below to explore its 3D semantic map.

šŸ—£ļø Roundtable Dialogues Core Feature 2

A generated conversation between cluster personas lets you hear contrasting themes as if the ideas were debating each other.

  • šŸ¤ One synthetic voice per cluster; tone & style guided by that cluster’s response vibe.
  • šŸ”„ Alternating lines trace tensions, agreements, and unique contributions.
  • šŸŽ§ Per‑line text‑to‑speech (distinct voices) + transcript highlighting while playing.
  • 🧪 Final Synthesis line summarizes convergence vs. unresolved friction.
  • āš™ļø Speed control, auto‑scroll, keyboard shortcuts (Space / ← / → / R).

Scroll below the 3D plot on a question page to access its dialogue module.

1,001
Survey Responses
16
Text Questions
3D
Semantic Space
AI-Powered
Embeddings
šŸŽØ AI & Art Perspectives
How do BC residents feel about AI's role in creative fields and artistic expression?
Loading…
Explore Creativity →
šŸ„ AI Helping Communities
Text responses about how AI can help British Columbia communities thrive.
Loading…
View Community Impact →
šŸŽ„ Community Video Insights
Video transcriptions about AI's potential to help BC communities.
Loading…
Watch Perspectives →
⚔ AI Superpowers
If AI had superpowers, how would British Columbians want them used for their communities?
Loading…
Explore Superpowers →
😰 AI Impact Worries
What concerns do BC residents have about AI's impact on society and their lives?
Loading…
Understand Concerns →
šŸŽ¬ Video Worries
Video transcriptions of British Columbians expressing their AI-related worries.
Loading…
Hear Concerns →
šŸ›ļø Advice to BC Leaders
What guidance do citizens want to give BC leaders about AI governance and policy?
Loading…
Read Advice →
šŸŽ™ļø Leadership Video Messages
Video transcriptions of advice and messages for BC's political leaders.
Loading…
Listen to Voices →
šŸ”® BC's AI Future
What matters most to British Columbians when thinking about the province's AI future?
Loading…
Shape the Future →
šŸ”ļø Indigenous AI Involvement
Perspectives on how Indigenous communities should be involved in BC's AI development.
Loading…
Honor Voices →
šŸŽ„ Indigenous Video Perspectives
Video insights about Indigenous communities and AI in British Columbia.
Loading…
Learn More →
āœ… Positive AI Sectors
Which sectors do BC residents think AI is making a positive impact in?
Loading…
See Positives →
āŒ Negative AI Sectors
Where do British Columbians see AI making negative impacts on society?
Loading…
Examine Negatives →
šŸ† Who Benefits from AI
Who do BC residents think gains the most from advances in artificial intelligence?
Loading…
Identify Winners →
šŸ›”ļø Most Trustworthy AI
What AI applications do British Columbians trust the most?
Loading…
Build Trust →
šŸ¦Ž AI Experience Stories
Personal descriptions of how BC residents relate to and experience AI technology.
Loading…
Share Experiences →

šŸ”¬ Methodology & Narrative Pipeline

The project blends unsupervised semantic mapping with generative narrative synthesis. Below is the end‑to‑end flow from raw free‑text to interactive 3D maps and voiced roundtable dialogues.

1. Ingestion & Cleaning: Open‑ended survey answers are de‑duplicated by exact text (frequency retained) and basic filters remove empty / placeholder values.

2. Semantic Embeddings: Each unique response is embedded into a 3,072‑dimensional vector (OpenAI text-embedding-3-large) capturing contextual meaning.

3. Dimensionality Reduction: 3D t‑SNE (adaptive perplexity) projects embeddings while preserving local semantic neighborhoods for visual exploration.

4. Automatic Clustering: K‑Means is run over the 3D projection for k ∈ [2,30]; the k with highest silhouette score is selected to represent emergent themes (no manual tuning).

5. Frequency & Sizing: Identical responses are aggregated; marker size scales by cube‑root(frequency) so common sentiments stand out without overwhelming rare but distinct ideas.

6. Color Mapping: A reproducible discrete palette assigns stable colors per cluster; the color legend is carried into both visualization and dialogue UI.

7. Cluster ā€œVibeā€ Extraction: For each cluster a one‑sentence tonal/style descriptor is LLM‑generated once and cached (e.g., ā€œpragmatic, solutions‑focused empathyā€).

8. Dialogue Transcript Generation: An LLM simulates a turn‑based exchange where each cluster speaks in its persona, referencing representative response content. A final Synthesis line summarizes convergence and tension points.

9. Text‑to‑Speech (Per Line): Each dialogue line is rendered to audio via distinct vetted voices (e.g., alloy, echo, fable, onyx…) with the vibe passed as system instructions (not spoken aloud).

10. Playback Assembly: A manifest ( audio_manifest.json) indexes lines, voices, cluster labels, and paths. The front‑end module streams sequential audio, highlights the active line, supports keyboard control, speed adjustment, preloading, and auto‑scroll.

11. Interactive Exploration: Users rotate / zoom 3D semantic maps, inspect original wording via hover, then scroll to the roundtable to ā€œhearā€ thematic perspectives converse—linking quantitative structure with qualitative nuance.

Reproducibility & Efficiency: Cluster vibes and transcripts are cached to avoid repeated LLM cost; visualization export + dialogue assets share a consolidated clustered_responses.json for downstream reuse.