<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> AVA is an AI-native Visual Analytics framework simplifying data analysis and visualization code generation. It targets engineers and researchers, shifting from rule-based to AI-driven capabilities, enabling natural language queries and LLM-powered insights for convenient data interaction.

How It Works

AVA uses a modular pipeline. Data ingestion supports CSV, JSON, URLs, and text. Its "smart data handling" automatically selects in-memory JavaScript for <10KB datasets or SQLite for larger ones. The analysis module generates code/SQL, with results summarized by an LLM into natural language responses, streamlining the visual analytics workflow.

Quick Start & Requirements

The framework can be initialized and used via TypeScript. A typical setup involves importing AVA and configuring LLM parameters such as model, apiKey, and baseURL. Data can be loaded using methods like loadCSV, loadObject, loadURL, or loadText. Subsequent analysis is performed via natural language queries. Prerequisites include a Node.js environment and valid LLM API credentials.

Highlighted Details

• Natural Language Queries: Users can pose questions about their data using plain English.
• LLM-Powered Analysis: Leverages large language models for intelligent data interpretation and insight generation.
• Smart Data Handling: Dynamically switches between JavaScript in-memory and SQLite based on data volume for optimal performance.
• Modular Architecture: Features a clean separation of concerns across data loading, analysis, and visualization modules.

Maintenance & Community

AVA is currently designated as an "experimental branch," welcoming contributions from the community. Related projects such as GPT-Vis, Chart Visualization Skills, and Vercel AI SDK are listed, indicating potential ecosystem integration.

Licensing & Compatibility

The project is released under the MIT License. This permissive license generally allows for broad use, including commercial applications and integration within closed-source projects, without significant restrictions.

Limitations & Caveats

As an experimental project, AVA has several features planned for future development. Notably, the visualization module, including chart recommendations and rendering integration, is not yet implemented. Support for streaming responses, advanced aggregation operations, and multi-table queries are also pending. Additional data source integrations beyond CSV, JSON, URL, and text are planned but not yet available.