Download
LIDA is an open-source library developed to automate the process of creating data visualizations and infographics using large language models. The system treats visualizations as executable code and uses AI to generate, modify, and interpret that code in order to transform raw datasets into meaningful charts and graphical explanations. Instead of requiring users to manually explore datasets and write plotting scripts, LIDA analyzes the data and automatically proposes visualization goals and design ideas that highlight patterns and relationships. The platform can generate visualization code compatible with a wide range of libraries, allowing it to integrate with common data science ecosystems. It also supports iterative workflows where visualizations can be edited, explained, evaluated, and repaired through AI-driven feedback loops. The system is model-agnostic and can connect to multiple language model providers, enabling flexibility across different AI infrastructures.
Features
- Automatic generation of data visualizations and infographics using large language models
- Data summarization module that converts datasets into natural language descriptions
- Goal generation system that proposes visualization objectives for exploratory analysis
- Visualization code generation compatible with multiple plotting libraries
- Tools for editing, explaining, evaluating, and repairing visualization code
- Support for multiple language model providers and flexible deployment environments
Project Samples
