AI Hackathon: AI Survey to Data Storytelling Pipeline

An end-to-end AI pipeline converts 500+ survey responses into a data-driven comic.
Recorded at the VanAI Round 1 Hackathon in Vancouver.

This presentation documents an AI workflow that transforms structured survey data into a narrative comic strip without manual illustration.
Key components:

Data ingestion – CSV import, cleansing, basic statistics.

Story synthesis – LLM-generated plot and dialogue from survey insights.

Image generation – Prompt-driven comparisons: ChatGPT-Vision vs. open-source diffusion models.

Automation layer – A CLI script accepts any edited CSV and outputs sequenced comic panels.

The GitHub repository contains the pipeline, prompt templates, and sample output. Vancouver AI researchers and creators can fork the code, replace the CSV, and produce new comics for meetups or demos.

Source: VanAI Round 1 Hackathon, Vancouver, BC.

Repository, slides, and dataset links are in the description.

#VanAI #VancouverAI #AIComicGeneration #DataStorytelling #GenerativeAI

📍 Chapters
0:00 Intro – VanAI Hackathon context
0:18 Project scope and objectives
0:45 Data cleaning: 500+ survey entries
1:14 Story synthesis with LLMs
1:40 Automated comic-generation pipeline
2:18 Panel walkthrough
3:05 Model comparison: ChatGPT vs open-source
3:35 Post-generation edits and typo fixes
4:06 Next steps for Vancouver AI community
4:41 Closing and repo link