https://github.com/kimtth/agent-data-analyst-stream-chainlit
⚡️Chainlit-based Data Analyst Chat Agent (Response API, Stream) 📈
https://github.com/kimtth/agent-data-analyst-stream-chainlit
agent chainlit code-interpreter data-analysis stream-response
Last synced: about 1 month ago
JSON representation
⚡️Chainlit-based Data Analyst Chat Agent (Response API, Stream) 📈
- Host: GitHub
- URL: https://github.com/kimtth/agent-data-analyst-stream-chainlit
- Owner: kimtth
- Created: 2025-09-04T05:41:43.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2025-09-04T08:25:47.000Z (about 1 month ago)
- Last Synced: 2025-09-04T10:31:19.941Z (about 1 month ago)
- Topics: agent, chainlit, code-interpreter, data-analysis, stream-response
- Language: Python
- Homepage:
- Size: 56.6 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 📊 Data Analyst Agent ⚡️
A small Chainlit-based data analyst chat agent that answers questions about a synthetic business dataset, runs pandas / matplotlib code when needed, and returns outputs (stream) and inline PNG plots.
## OpenAI Code Interpreter
- Released in **March 2023** (alpha for ChatGPT Plus).
- Integrated with **Assistant API**.
- **Assistant API** will be deprecated in 2026 ⚠️; use **Responses API** instead.
- **Responses API** creates an isolated container to execute Python code, with outputs such as images stored within the container.
- **Open-source alternatives** (e.g., Open Interpreter) are not actively developed; using **Responses API** is recommended. Integration with **Chainlit** has issues:
1. [Open Interpreter](https://github.com/OpenInterpreter/open-interpreter) (v0.4.3) sometimes fails to install with Chainlit (v2.7.2) due to dependency collisions.
2. [Code Interpreter API](https://github.com/shroominic/codeinterpreter-api) (v0.1.20) has issues with Azure OpenAI.
## Synthetic data description
- **synthetic\_business\_data.csv**
- Synthetic, tabular business / transaction data for demo and analysis.
- **synthetic\_metadata.json**
- JSON object describing columns and their human-readable descriptions.
## Usage
1. Create a `.env` with required Azure variables:
- `AZURE_OPENAI_ENDPOINT`
- `AZURE_OPENAI_DEPLOYMENT_NAME`
- `AZURE_OPENAI_API_KEY`
- `AZURE_OPENAI_API_VERSION`
2. Install dependencies:
- `poetry install`
3. Run the app:
- `python app.py`
4. In the chat, ask questions about the dataset. The assistant will run code for calculations/plots and return results.
## Notes
- Tool execution writes a Python script, runs it, and returns stdout/stderr plus any generated PNGs inline.
- File uploads are disabled in the chat flow.
- `app.legacy.py` is present in the repo but is considered legacy / not used by the code.
- If you plan to deploy or extend for Azure, follow Azure best practices, secure `.env` secrets 🔐, and review deployment guidance.