Maritime Navigator 🚢

Maritime Navigator is an application designed to predict maritime traffic conditions on specific routes, given weather conditions and the desired date. It utilizes TabNet, a deep learning architecture, to analyze data and generate predictions. The application bridges its backend and frontend seamlessly using Gradio, enabling easy interaction and visualization.

---

Features ✨

• Predicts maritime traffic for specific routes on a given day.
• Incorporates weather conditions into predictions.
• Uses the TabNet architecture for accurate and efficient deep learning.
• Interactive Gradio interface for user-friendly predictions.
• Fully hosted and runnable on Google Colab.

---

How It Works 🔍

The workflow involves two stages:

1. Training the Model:

   • A TabNet deep learning model is trained on maritime data to learn traffic patterns under various weather conditions.

2. Making Predictions:

   • The trained model predicts maritime traffic for a given route and date using user-provided data.

---

Prerequisites 🛠️

• A Google account to access Google Colab.
• Basic familiarity with running code in Google Colab.

---

How to Run the Application 🚀

Step 1: Train the Model

1. Open the following Colab notebook:
   Train the Model Notebook

2. Run each block of code sequentially:

   • Follow the prompts and ensure that all dependencies are installed.
   • At the end of execution, a trained model file (model.pt) will be downloaded to your computer.

Step 2: Use the Trained Model

1. Open the second Colab notebook:
   Prediction Notebook

2. Run each block of code sequentially:

   • When prompted, upload the trained model file downloaded earlier (model.pt).
   • Provide the required data inputs (route, date, and weather conditions).
   • The code will process the inputs and display the maritime traffic predictions.

---

Experimentation ⚡

Feel free to experiment with different values in the dataset provided in the Prediction Notebook to observe how traffic predictions vary with input parameters.

---

Technologies Used 🖥️

• TabNet: Deep learning architecture for tabular data.
• Gradio: Frontend framework for linking machine learning models to user interfaces.
• Google Colab: Cloud-based environment for running Python code.

---

Future Improvements 🚀

• Expand prediction capabilities to include multiple routes simultaneously.
• Integrate live weather data from external APIs for real-time predictions.
• Deploy the application as a standalone web or mobile app.

---

Screenshots 📸