PandasAI

PandasAI is a Python library that makes it easy to ask questions to your data in natural language. It supports interacting with databases or datalakes (SQL, CSV, parquet) and makes data analysis conversational using LLMs and RAG.

Main Features:

1. Ask questions: Users can ask questions to their data in natural language, such as "What is the average revenue by region?" or "What is the total sales for the top 3 countries by sales?".
2. Visualize charts: Supports generating charts via natural language prompts, for instance, plotting a histogram of countries showing the gdp with different colors for each bar.
3. Multiple DataFrames: Allows passing in multiple dataframes and asking questions relating them, such as combining employees and salaries data to ask "Who gets paid the most?".
4. Docker Sandbox: Provides a secure, isolated environment to execute code safely and mitigate the risk of malicious attacks. By installing the pandasai-docker package, users can initialize a DockerSandbox for secure chat interactions and must stop the sandbox when done.

Usage Instructions:

• Python Requirements: Version 3.8+ <=3.11.
• Installation: Can be installed using pip or poetry for the core library pandasai and LLM integrations like pandasai-litellm.
• Configuration and Execution: Requires initializing an LLM (e.g., gpt-4.1-mini) using LiteLLM, configuring PandasAI with pai.config.set, loading data via pai.read_csv or pai.DataFrame, and calling the .chat() method or pai.chat() function.

Target Users: Non-technical users looking to interact with their data in a more natural way, and technical users looking to save time and effort when working with data.

Core Advantages: Transforms complex data querying and analysis into simple natural language conversations; supports multiple data source formats; ensures code execution security via Docker Sandbox; offers great flexibility as an open-source library.

Pricing Information: The PandasAI library is available under the MIT expat license (except for the pandasai/ee directory which has its own license). For managed PandasAI Cloud or self-hosted Enterprise Offering, users need to contact the team for pricing details.