lida

What it is

LIDA is a library for generating data visualizations and data-faithful infographics. LIDA is grammar agnostic (will work with any programming language and visualization libraries e.g. matplotlib, seaborn, altair, d3 etc) and works with multiple large language model providers (OpenAI, Azure OpenAI, PaLM, Cohere, Huggingface). Details on the components of LIDA are described in the paper here and in this tutorial notebook. See the project page here for updates!.

Note on Code Execution: To create visualizations, LIDA generates and executes code. Ensure that you run LIDA in a secure environment.

How it works

• LIDA treats visualizations as code and provides an API for generating, executing, editing, explaining, evaluating and repairing visualization code.
• Features include: Data Summarization, Goal Generation, Visualization Generation, Visualization Editing, Visualization Explanation, Visualization Evaluation and Repair, Visualization Recommendation, Infographic Generation (beta).
• It can interface with multiple providers via llmx and supports a CLI/web UI.

Getting started

• Setup requires Python 3.10 or higher. Commands shown:

pip install -U lida 

pip install -U llmx openai

export OPENAI_API_KEY=<your key>

• LIDA provides a bundled UI and Web API that can be started with:

lida ui  --port=8080 --docs

• Web API and UI can also be built with Docker using:

docker compose up

• If you prefer dev installation from source:

pip install -e .

Recent releases

• No releases listed in RELEASES (latest 0).

Traction

• GitHub repo shows 3273 stars, 381 forks, 88 open issues as of the provided data.

Caveats

• License: MIT.
• Important notes include: LIDA generates and executes code; best used with datasets having a small number of columns (<= 10); secure environment recommended; requires appropriate permissions for code execution.
• Documentation references external resources and example notebooks, including a tutorial notebook and arXiv paper.