Back to Projects

ChartAgent Demo

Turn questions into visualizations instantly

Interactive AI dashboard with Text-to-Chart capabilities via conversational prompts and code generation.

The Problem

Data teams spend hours writing SQL and building charts. Business users wait days for simple dashboards. The gap between "I want to see X" and actually seeing X is way too wide. Existing BI tools require training, SQL knowledge, or both. Self-service analytics promised to solve this but ended up creating another tool people need to learn. What if you could just ask questions in plain English and get instant visualizations?

What You Can Do

Ask in Plain English

Type 'Show me sales by region for Q4' and get an instant chart.

Generate Charts Instantly

The AI writes the code, generates the chart, and explains what it did.

Iterate Conversationally

Refine your visualization with follow-up questions: 'Now group by category'.

Export Anywhere

Download charts as images or export the underlying data for further analysis.

Tech Stack

AI & LLM

OpenAI APIAnthropic APICode Generation

Multi-provider LLM support for natural language to code translation.

Data Processing

DuckDBPandasNumPy

Lightning-fast in-memory analytics for instant results.

Visualization

PlotlyInteractive ChartsExport APIs

Rich interactive visualizations with export capabilities.

Interface

StreamlitReal-time UpdatesSession State

Clean, intuitive UI that makes data exploration effortless.

Architecture

Simple but powerful architecture:

1. Chat Interface: Streamlit-based conversational UI

2. LLM Engine: Translates questions to Pandas/DuckDB queries

3. Execution Sandbox: Safely runs generated code

4. Renderer: Plotly visualization with interactive features

The agent maintains conversation context, so you can build on previous queries and refine visualizations iteratively.

Results

< 5s
Question to visualization
Zero
SQL knowledge required
Multi-LLM
OpenAI + Anthropic support

Key Features

  • 1Natural language to chart generation
  • 2Interactive Plotly visualizations
  • 3DuckDB for fast in-memory analytics
  • 4Support for multiple LLM providers
  • 5Streamlit-based intuitive interface

Interested in this project?

Check out the source code, try the demo, or get in touch to discuss how similar solutions could help your team.