OceanPilot AI: AI-Driven Marine Data Platform

OceanPilot AI is a web-based, cloud-ready platform designed to solve a critical data challenge for the Centre for Marine Living Resources & Ecology (CMLRE). It ingests, processes, and analyzes complex, heterogeneous marine datasets, turning siloed information into actionable, cross-disciplinary insights for scientists and policymakers.

This project was developed for the Smart India Hackathon 2025.

Demo of Data Analysis

The Problem

Marine scientists at CMLRE work with diverse datasets—oceanography, fisheries, and molecular biology—that are stored in different formats and siloed systems. This fragmentation makes integrated analysis, real-time visualization, and advanced analytics nearly impossible, creating a major bottleneck for critical research on ecosystem conservation and fisheries management

The Solution

OceanPilot AI provides a unified platform that automates the entire data lifecycle. It leverages an LLM-powered assistant to make complex data interaction intuitive for non-expert users, enabling deep analysis and visualization through natural language.

Key Features

• Automated Data Ingestion & Preprocessing: Users can upload diverse raw data files (e.g., .gz, .csv). The platform automatically processes them, provides a data quality summary, and handles null values, preparing the data for analysis.

• LLM-Powered Co-Pilot: An integrated AI assistant helps users at every step, from providing automated data summaries to suggesting intelligent schema mappings for new datasets.

• Dynamic, Agent-Driven Analysis: Users can query the data using natural language or select pre-defined analyses. An AI agent interprets these requests to generate complex visualizations on-demand, such as Temperature vs. Depth profiles and T-S (Temperature-Salinity) diagrams.

Technical Architecture

The platform is built on a modern, scalable tech stack designed for real-time data processing and interactive user experiences.

(High-level workflow of the OceanPilot AI platform)

Tech Stack

• Backend: FastAPI , LangChain (for LLM orchestration and agents) , Pandas (for data handling).
• Frontend: React.js (for a dynamic and responsive UI).
• Database: Supabase (for storing datasets and session metadata).
• Real-time Communication: WebSockets (for live updates between the frontend and backend during analysis).

Setup & Installation

To run this project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/IOutis/OceanPilotAI.git

2. Env File Structure: Create a .env file

   GEMINI_API_KEY=*********
   

3. Set up the backend:

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt
   uvicorn backend.main:app --reload