Published:  ·  4:45

In this demo, Prajwal Prashanth showcases his Round 3 Vancouver AI Hackathon submission: a storytelling tool that turns open-ended survey data into living conversations.

Using semantic embeddings, 3D maps, and persona-based roundtables, this project analyzes 1,001 open-ended responses from British Columbians about AI’s future—covering creativity, ethics, governance, Indigenous involvement, and more.

✨ Features:

Semantic Maps: Visualize how ideas cluster in 3D space, from consensus hubs to edge perspectives.

Persona Roundtables: Each cluster is voiced as a “representative speaker,” debating and dialoguing in real time.

Empathetic Storytelling: Synthesis highlights agreements without erasing disagreements, making data human-centered and accessible.

🔗 Try the Demo: https://prajwal.is-a.dev/vanai3/bc_ai_hackathon_round_3

💻 Repo: https://github.com/Toricane/vanai3

📡 From raw responses to plural voices—this project reframes survey analysis as storytelling across difference.

👉 Subscribe for more on AI, data storytelling, and community innovation.

Prajwal takes 1,001 raw survey voices from British Columbia and transforms them into a two-stage storytelling interface: first, a 3D semantic map where related responses cluster into visible constellations of meaning; second, a generated roundtable discussion where each cluster is voiced as a persona in conversation.

The result is a dialogue between “consensus hubs” and “edge perspectives” that surfaces agreements, tensions, and trade-offs without flattening dissent. Instead of scrolling lines of comments or collapsing nuance into a bar chart, users can rotate the map, hover for authentic text, then listen as the clusters speak to each other in a staged, turn-based debate.

A final synthesis voice draws threads together while preserving disagreement. It’s data storytelling as theatre — plural, embodied, and unforgettable.
——–

A dual-interface exploration tool:

– 3D Semantic Maps → plot open-text responses into natural clusters.
– Persona Roundtable → each cluster speaks in its own voice, debating others.

Together, they give both structural clarity (map of ideas) and empathetic clarity (voices in dialogue).

Watch on YouTube →