title	AI Data Analyst
emoji	⚡
colorFrom	blue
colorTo	indigo
sdk	docker
pinned	false
license	mit

📊 AI Data Analyst

An AI-powered data analyst built with FastAPI and Google Gemini API. It accepts natural language questions and supporting files, then returns structured JSON answers, including data analysis and charts encoded as base64 images.

This app is designed for:

• 📊 Automated data analysis
• 📈 Chart generation (base64 PNGs)

🚀 Live Deployment

You can checkout the app by clicking the below image link.

⚙️ Features

• ⚡ FastAPI async server – Handles multiple requests concurrently with high performance.
• 📝 Per-request server logging – Detailed logs for each request with request ID, client info, execution status, and timing.
• 🧠 Powered by Gemini – Leverages Google’s latest LLM for reasoning, analysis, and code generation.
• 🔀 Load balancing – Rotates across multiple Gemini API keys for improved stability and performance.
• 📂 File handling – Accepts both text (e.g., CSV, TXT, Markdown) and binary (e.g., images, PDFs) files.
• 🤖 Dynamic code generation – LLM generates Python code dynamically to answer questions.
• 🔄 Automatic error correction – Re-prompts Gemini to fix code when execution errors occur.

🛠️ Local Setup

1. Clone the repository

   git clone git@github.com:lightRajat/ai-data-analyst.git
   cd ai-data-analyst

2. Install dependencies

   uv sync

   Note: If you don't have uv installed, you can install it using curl -LsSf https://astral.sh/uv/install.sh | sh.

3. Store your Gemini API keys inside .env file separated by commas.

   GEMINI_API_KEYS="key1,key2,key3"

4. Run the app

   uv run uvicorn main:app --host 0.0.0.0 --port 7860 --reload

📤 Example Usage

Let's say you want to analyze a network interaction graph to find the most connected individual and visualize the network structure. You have your data in a file named edges.csv.

How to use the Web Dashboard

1. Upload Data: Drag and drop your edges.csv file into the Attachments zone.

2. Define Analysis Task: You need to tell the AI what to do. You can type this directly into the Analysis Plan text box or drag and drop a text file (e.g., analysis.txt).

   Important: You must explicitly define the JSON response structure in your request. This tells the AI exactly what format to return so the dashboard can render charts and data correctly.

3. Run Analysis

Example analysis.txt (Your Plan)

Use the undirected network in `edges.csv`.

Return a JSON object with keys:
- `edge_count`: number
- `highest_degree_node`: string
- `average_degree`: number
- `density`: number
- `shortest_path_alice_eve`: number
- `network_graph`: base64 PNG string under 100kB
- `degree_histogram`: base64 PNG string under 100kB

Example edges.csv (Your Attachment)

source,target
Alice,Bob
Alice,Carol
Bob,Carol
Bob,David
Bob,Eve
Carol,David
David,Eve

📜 License

This project is licensed under the MIT License – see the LICENSE file for details.