PulseLab

About The Project

PulseLab is a comprehensive data science project focused on forecasting the price of Bitcoin (BTC-USD). This repository explores, implements, and evaluates a wide array of time-series forecasting models, ranging from classical statistical methods to advanced deep learning and gradient-boosting ensembles.

The primary goal is to identify the most accurate and reliable model for short-term price prediction by comparing their performance on a hold-out test set. The project culminates in a demonstration UI that showcases the predictive power of the top-performing models.

Models Implemented

This analysis provides a comparative leaderboard for the following models:

• Statistical Models:
  • ARIMA (Autoregressive Integrated Moving Average)
  • SARIMA (Seasonal ARIMA)
  • Holt-Winters
• Machine Learning Models:
  • LightGBM (Light Gradient Boosting Machine)
  • XGBoost
  • CatBoost
  • Random Forest
• Deep Learning Models:
  • ANN (Artificial Neural Network - MLP)
  • LSTM (Long Short-Term Memory)
  • SimpleRNN

Based on the analysis, LightGBM was identified as the top-recommended model due to its superior accuracy (lowest MAPE) and excellent generalization, showing no signs of overfitting.

Project Structure

PulseLab/
├── BTC-USD_2022-06-30_to_2025-09-30.csv  # The primary dataset used for training and testing
├── Capstone_Finalized.ipynb             # Main Jupyter Notebook with all model implementations
├── Demo_Capstone_UI_+_Tree_Based_Models.ipynb # Notebook for the demo UI (likely Streamlit/Gradio)
├── requirements.txt                     # All Python dependencies needed to run the project
├── LICENSE                                # MIT License
└── README.md                              # You are here!

Getting Started

To get a local copy up and running, follow these simple steps.

Prerequisites

• Python 3.8 or higher
• Jupyter Notebook or Jupyter Lab or colab works too!

Installation

1. Clone the repo

   git clone https://github.com/SmridhVarma/PulseLab.git

2. Navigate to the project directory

   cd PulseLab

3. Install the required packages

   pip install -r requirements.txt

Running the Project

You can explore the complete analysis or run the UI demo - refer to the demo vid to get a feel of the end product.

1. To see the full analysis and model comparisons:

   • Launch Jupyter Notebook:

     jupyter notebook

   • Open Capstone_Finalized.ipynb.

2. To run the UI demonstration:

   • Launch Jupyter Notebook:

     jupyter notebook

   • Open Demo_Capstone_UI_+_Tree_Based_Models.ipynb and run the cells.

License

Distributed under the MIT License. See LICENSE for more information.