Aviation Prediction - FastAPI Backend & Next.js Frontend

Link to over video presentation

Video Presentation

Link to Google Drive

• Consists the whole code with charts and datasets. Google Drive

Overview

This repository contains a FastAPI backend that allows users to train a machine learning model with custom hyperparameters and predict airline profits based on provided data. Additionally, it includes a Next.js frontend for user interaction, a Jupyter Notebook demonstrating the model's performance using XGBRegressor, and an Exploratory Data Analysis (EDA) of the dataset.

Project Background

This project was developed for a hackathon, where the challenge was to train a model to predict airline profits. Our primary goal was to develop a highly accurate machine learning model. To enhance usability, we also built a web application that enables users to:

• Tune hyperparameters dynamically and visualize their impact.
• Input their own data and obtain profit predictions using a pre-trained model.
• Explore detailed EDA insights and a model performance notebook.

Tech Stack

• Backend: FastAPI (Located in backend/ folder)
• Frontend: Next.js (Located in dashboard/ folder)
• Model: XGBRegressor (Extreme Gradient Boosting Regression)
• EDA & Model Notebook: Jupyter Notebook (Located in the root directory)
• Deployment: Next.js frontend deployed (Specify hosting if applicable)

Features

1. Machine Learning Model

• Uses XGBRegressor for high-accuracy predictions.
• Trained with custom hyperparameters.
• Evaluated using metrics like RMSE and R².
• Results documented in Model.ipynb.

2. FastAPI Backend

• API endpoints for:
  • Training a model with custom hyperparameters
  • Predicting airline profits based on user-input data
  • Generating visualizations for hyperparameter tuning
• Located in backend/ folder.

3. Next.js Frontend (Dashboard)

• Interactive UI to:
  • Upload datasets
  • Tune hyperparameters
  • View real-time predictions and performance metrics
• Located in dashboard/ folder.

4. Jupyter Notebook - Model & EDA

• Exploratory Data Analysis (EDA):
  • Insights into airline profit data
  • Feature importance analysis
  • Correlation between variables
• Model Training Notebook:
  • Training process using XGBRegressor
  • Accuracy metrics, RMSE, R² score, etc.
  • Comparisons with baseline models

Installation & Setup

Prerequisites

Ensure you have Python 3.11 and Node.js installed.

Clone the repository

git clone https://github.com/Bhawesh-Agrawal/NEURAL-NINJAS.git
cd your-repo

Backend Setup

cd backend
pip install -r requirements.txt
uvicorn main:app --reload

It will only work with python 3.11

Server will be running at: http://127.0.0.1:8000

Access the API Documentation

FastAPI provides an interactive API documentation at:

• Swagger UI: http://127.0.0.1:8000/docs
• Redoc: http://127.0.0.1:8000/redoc

Frontend Setup (Next.js)

cd dashboard
npm install
npm run dev

Front end will be available at http://localhost:3000

Running the Jupyter Notebook

Run it in the root folder of the project

pip install -r requirements.txt
jupyter notebook

We have custom function after running all cells you will be promted to choode from 2 option either you can upload a csv file or input a array and the model will predict the output for you.

Open and explore EDA.ipynb and Model

API Endpoints

Method	Endpoint	Description
POST	/train	Train the model with custom hyperparameters
POST	/predict	Predict airline profit using input features

Future Enhancements

• Optimize model hyper parameters for better accuracy
• Deploy the model on a cloud server
• Improve front end UX with better visualizations
• Implement additional ML models for comparison

Contributors

Neural Ninjas

• Bhawesh Agrawal
• Priyanka Singh
• Shasank Sachan

---