Spaces:
Sleeping
Sleeping
Upload README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: LLM Data Analyst Agent
|
| 3 |
+
emoji: 🤖
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: green
|
| 6 |
+
sdk: streamlit
|
| 7 |
+
sdk_version: 1.32.0
|
| 8 |
+
app_file: app.py
|
| 9 |
+
pinned: false
|
| 10 |
+
license: apache-2.0
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# 🤖 LLM-powered Data Analyst Agent
|
| 14 |
+
|
| 15 |
+
An intelligent data analysis assistant that helps you explore and understand customer support datasets using advanced language models.
|
| 16 |
+
|
| 17 |
+
## 🌟 Features
|
| 18 |
+
|
| 19 |
+
- **Interactive Data Analysis**: Ask questions in natural language and get intelligent responses
|
| 20 |
+
- **Multiple Planning Modes**: Choose between pre-planning and reactive dynamic planning
|
| 21 |
+
- **Beautiful UI**: Modern, responsive interface with custom styling
|
| 22 |
+
- **Real-time Conversations**: Chat-like interface for seamless interaction
|
| 23 |
+
- **Dataset Insights**: Automatic analysis of customer support conversations
|
| 24 |
+
|
| 25 |
+
## 🚀 How to Use
|
| 26 |
+
|
| 27 |
+
1. **Ask Questions**: Type your question about the customer support data
|
| 28 |
+
2. **Get Insights**: The AI will analyze the data and provide detailed answers
|
| 29 |
+
3. **Explore Further**: Follow up with additional questions for deeper analysis
|
| 30 |
+
|
| 31 |
+
### Example Questions:
|
| 32 |
+
- "What are the most common customer issues?"
|
| 33 |
+
- "Show me examples of billing problems"
|
| 34 |
+
- "What's the distribution of customer intents?"
|
| 35 |
+
- "Summarize the main categories of support requests"
|
| 36 |
+
|
| 37 |
+
## 🛠️ Technology Stack
|
| 38 |
+
|
| 39 |
+
- **Frontend**: Streamlit with custom CSS styling
|
| 40 |
+
- **AI Model**: Nebius API (Qwen/Qwen3-30B-A3B)
|
| 41 |
+
- **Data Processing**: Pandas for data manipulation
|
| 42 |
+
- **Dataset**: Bitext Customer Support Dataset
|
| 43 |
+
|
| 44 |
+
## 📊 Dataset
|
| 45 |
+
|
| 46 |
+
This app analyzes the [Bitext Customer Support Dataset](https://huggingface.co/datasets/bitext/Bitext-customer-support-llm-chatbot-training-dataset) which contains real customer support conversations with:
|
| 47 |
+
|
| 48 |
+
- **Categories**: Different types of customer issues
|
| 49 |
+
- **Intents**: Specific customer intentions
|
| 50 |
+
- **Customer Messages**: Original customer inquiries
|
| 51 |
+
- **Agent Responses**: Support agent replies
|
| 52 |
+
|
| 53 |
+
## 🔧 Configuration
|
| 54 |
+
|
| 55 |
+
The app requires a Nebius API key to function. This has been configured as an environment variable for this Space.
|
| 56 |
+
|
| 57 |
+
## 💡 Tips
|
| 58 |
+
|
| 59 |
+
- **Be Specific**: More specific questions often yield better insights
|
| 60 |
+
- **Explore Different Angles**: Try both quantitative ("how many") and qualitative ("why") questions
|
| 61 |
+
- **Use Follow-ups**: Build on previous answers for deeper analysis
|
| 62 |
+
|
| 63 |
+
## 🎯 Planning Modes
|
| 64 |
+
|
| 65 |
+
- **Pre-planning**: The agent first classifies your question, then executes analysis
|
| 66 |
+
- **Reactive Planning**: The agent dynamically decides how to approach your question
|
| 67 |
+
|
| 68 |
+
Choose the mode that works best for your analysis style!
|
| 69 |
+
|
| 70 |
+
---
|
| 71 |
+
|
| 72 |
+
*Built with ❤️ using Streamlit and powered by advanced language models*
|