Time Series Analysis and Forecasting

This repository contains a machine learning project for time series analysis and forecasting using multiple models such as Linear Regression, ARIMA, SARIMA, and PROPHET. The project is implemented in Python and includes Jupyter notebooks for exploratory data analysis (EDA) and a main script (main.py) to run the forecasting analysis.

Table of Contents

• Time Series Analysis and Forecasting
  • Table of Contents
  • Installation
    • 1. Clone the Repository
    • 2. Create a Virtual Environment
    • 3. Install Dependencies
  • Project Structure
  • Usage
  • Exploratory Data Analysis (EDA)
  • Dependencies
  • License

Installation

1. Clone the Repository

git clone https://github.com/TJselevani/python-analysis.git
cd python-analysis

2. Create a Virtual Environment

python3 -m venv venv
source venv/bin/activate  # On macOS/Linux
venv\Scripts\activate    # On Windows

3. Install Dependencies

pip install -r requirements.txt

Project Structure

├── config.py                 # Configuration file
├── data/                     # Directory containing raw data
├── eda/                      # Jupyter notebooks for EDA
│   ├── day.ipynb
│   ├── week.ipynb
│   ├── month.ipynb
│   ├── year.ipynb
├── files/                    # Directory for storing analysis files
├── main.py                    # Main script for running forecasting
├── main.ipynb                 # Jupyter notebook version of the main script
├── utils/                     # Helper functions
│   ├── data_preparation/
│   │   ├── prepare_time_series.py
│   ├── ensemble/
│   │   ├── ensemble_forecast.py
│   │   ├── plot_ensemble_forecast.py
│   ├── prediction/
│   │   ├── ensemble_prediction.py
├── requirements.txt           # List of dependencies
├── README.md                  # Project documentation

Usage

To run the time series forecasting analysis, execute the main.py script:

python main.py

This script:

1. Loads the dataset from config.py.
2. Runs predictions using different models.
3. Creates an ensemble forecast.
4. Visualizes and prints the forecast for the next 7 days.

Exploratory Data Analysis (EDA)

The eda/ directory contains Jupyter notebooks for data visualization:

• day.ipynb: Analysis at the daily level.
• week.ipynb: Weekly trends.
• month.ipynb: Monthly patterns.
• year.ipynb: Yearly trends.

To run the EDA notebooks:

jupyter notebook

Then open the desired notebook from the eda/ directory.

Dependencies

A list of required dependencies is available in requirements.txt. The major dependencies include:

• pandas
• numpy
• matplotlib
• scikit-learn
• statsmodels
• prophet
• plotly
• jupyter
• fastapi
• uvicorn

License

This project is open-source and available under the MIT License.

---

For any issues or contributions, feel free to submit a pull request or open an issue in the repository!