|
|
--- |
|
|
title: Pandas CSV Analyzer (LlamaIndex) |
|
|
emoji: 🐼 |
|
|
colorFrom: purple |
|
|
colorTo: indigo |
|
|
sdk: gradio |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
license: mit |
|
|
short_description: An AI agent for CSV analysis with LlamaIndex & Pandas. |
|
|
sdk_version: 5.44.1 |
|
|
--- |
|
|
|
|
|
# 🐼 Pandas CSV Analyzer (LlamaIndex) |
|
|
|
|
|
This is an advanced AI agent that allows you to analyze CSV files by asking questions in natural language. Upload a file, ask your questions, and get instant insights without writing a single line of code. |
|
|
|
|
|
## ✨ Key Features |
|
|
|
|
|
* **Natural Language Queries:** Ask questions like, "Which branch had the highest total revenue?" or "Show the top 5 best-selling products." |
|
|
* **Pandas Code Generation:** The AI generates and executes the necessary Python (Pandas) code to answer your query, displaying the code used for full transparency. |
|
|
* **PDF Report Generation:** Download your entire conversation history and analysis in a clean and organized PDF report. |
|
|
* **Robust Architecture:** Built with the new LlamaIndex Workflows library, ensuring a modular and reliable data processing pipeline. |
|
|
|
|
|
## 🚀 How to Use |
|
|
|
|
|
1. **Upload your CSV File:** Use the upload panel on the left to load your data. |
|
|
2. **Ask a Question:** Type your question about the data in the text box at the bottom. |
|
|
3. **Get Insights:** The agent will process your request, generate the answer, and display it in the chat. |
|
|
4. **Download the Report:** When your analysis is complete, click "Generate and Download PDF" to get a full report. |
|
|
|
|
|
## 🛠️ How it Works (Tech Stack) |
|
|
|
|
|
This project integrates several cutting-edge technologies to create a seamless data analysis experience: |
|
|
|
|
|
* **Interface:** **Gradio** is used to build the interactive web interface. |
|
|
* **Data Manipulation:** **Pandas** is the engine behind all data manipulation and analysis of the CSV file. |
|
|
* **AI Orchestration:** **LlamaIndex Workflows** manages the entire process, from receiving the user's question to synthesizing the final answer. This is a modern replacement for the older `QueryPipelines`. |
|
|
* **Language Model (LLM):** The **Groq API** provides access to high-speed language models (like Llama 3) to generate Pandas code and synthesize human-readable answers. |
|
|
* **PDF Generation:** The **FPDF2** library is used to create reports from the conversation history. |
|
|
|
|
|
The workflow is as follows: |
|
|
1. The user submits a question. |
|
|
2. The LlamaIndex `PandasWorkflow` is initiated. |
|
|
3. The LLM receives the question, data schema, and examples, then generates a Pandas expression. |
|
|
4. This expression is safely executed in the backend. |
|
|
5. The result is sent back to the LLM, which generates a final, natural-language response for the user. |
|
|
|
|
|
--- |
|
|
|
|
|
*Developed by Yuri Arduino Bernardineli Alves* |
|
|
* **GitHub:** [YuriArduino](https://github.com/YuriArduino) |
|
|
* **Email:** yuriarduino@gmail.com |
