Himalayan Expedition Success Prediction

A machine learning project for predicting the success of Himalayan expeditions using historical data.

Project Overview

This project analyzes Himalayan expedition data to predict expedition success using various machine learning models including Logistic Regression, KNN, Decision Trees, Neural Networks, Balanced Random Forest, and Easy Ensemble AdaBoost.

Project Structure

├── dataset/                 # Raw CSV datasets
│   ├── exped.csv
│   ├── members.csv
│   └── peaks.csv
├── Models/
│   ├── data/               # Processed data
│   ├── frontend/           # Streamlit web application
│   ├── models/             # Trained model files (.h5, .pkl)
│   ├── notebooks/          # Jupyter notebooks for each model
│   ├── utils/              # Utility functions
│   ├── config.py           # Configuration settings
│   ├── prepare_data.py     # Data preprocessing script
│   └── requirements.txt    # Python dependencies

Features

• Multiple ML models for expedition success prediction
• Interactive web application built with Streamlit
• Data preprocessing and feature engineering
• Model comparison and evaluation
• Real-time prediction with confidence scores
• Visual model performance comparison

Installation

1. Clone the repository
2. Install dependencies:

pip install -r Models/requirements.txt

Usage

Data Preparation

cd Models
python prepare_data.py

Train Models

Open and run the Jupyter notebooks in Models/notebooks/ to train models:

1. Train each model (Logistic Regression, KNN, etc.)
2. Models are automatically saved to Models/models/

Run Web Application

Locally:

streamlit run Models/frontend/app.py

On Streamlit Cloud:

1. Push code to GitHub
2. Deploy at streamlit.io/cloud
3. Main file: Models/frontend/app.py

See DEPLOYMENT.md for detailed deployment instructions.

Application Screenshots

Step 1: Dashboard Overview

The main dashboard displays available models and their current status.

Step 2: Model Accuracy Comparison

View and compare the accuracy of all trained models.

Step 3: Make Predictions

Enter expedition parameters (age, team size, season, peak, etc.) to predict success probability.

Step 4: View Prediction Results

Get instant predictions showing whether the expedition is likely to succeed or fail.

Step 5: Confidence Scores

See confidence scores from all models for comprehensive analysis.

Models Implemented

1. Logistic Regression
2. K-Nearest Neighbors (KNN)
3. Decision Tree
4. Neural Network
5. Balanced Random Forest
6. Easy Ensemble AdaBoost

Technologies Used

• Python 3.x
• Machine Learning: scikit-learn, TensorFlow, imbalanced-learn
• Data Analysis: pandas, numpy
• Visualization: matplotlib, seaborn, plotly
• Web App: Streamlit

Dataset

The project uses Himalayan expedition data including:

• Expedition details
• Member information
• Peak characteristics

License

This project is for educational purposes.