Introduction

DataViz Storyteller

DataViz Storyteller is a conversational data analysis application. Upload datasets, ask questions in natural language, and get summaries, visualizations, SQL, and analysis directions aligned with Machine Learning, Data Analytics, and Data Visualization best practices.

What you can do

• Analysis and runnable code: Built-in prompts (type / in chat) guide you through data quality, EDA, descriptive/diagnostic/predictive analytics, and visualization. The AI provides runnable Python or SQL in fenced code blocks; run them in the notebook with the Run button (dataset is available as df).
• Charts and summaries: Generate bar, line, scatter, box, heatmap, and other charts plus data summaries inside the chat.
• Notebook panel: The right panel shows the thread as notebook-style cells (In/Out), or embeds the real Jupyter notebook when NEXT_PUBLIC_JUPYTER_LAB_URL is set.
• Natural language: Filter, group, run SQL, summarize, and plot by describing what you want.

Concepts behind the app

The app steers analysis using ML, DA, and DV concepts: supervised/unsupervised learning, descriptive/diagnostic/predictive analytics, and visualization principles. See Concepts (ML, DA, DV) for the full overview.

Tech stack

• Next.js — app and API routes
• Tambo — chat state, components, and tools (100% open source: tambo-ai/tambo)
• Jupyter MCP (optional) — real notebook in the panel; the AI writes and runs cells via MCP

Idea and feature credits go to Claude.