Time Series Trend Analysis

A comprehensive Python toolkit for time series trend analysis, forecasting, and anomaly detection. This project provides both traditional statistical methods and modern machine learning approaches for analyzing time series data.

Features

• Trend Analysis: Moving averages, polynomial regression, and detrending
• Anomaly Detection: Isolation Forest and statistical methods
• Forecasting: ARIMA, Prophet, LSTM, and Linear Regression
• Visualization: Comprehensive plotting and dashboard creation
• Web Interface: Interactive Streamlit application
• CLI Interface: Command-line tool for batch processing
• Configuration: YAML-based configuration system
• Testing: Comprehensive unit test suite

Installation

1. Clone the repository:

git clone https://github.com/kryptologyst/Time-Series-Trend-Analysis.git
cd Time-Series-Trend-Analysis

2. Install dependencies:

pip install -r requirements.txt

3. (Optional) Install additional dependencies for specific features:

# For Prophet forecasting
pip install prophet

# For LSTM forecasting
pip install torch

# For ARIMA forecasting
pip install statsmodels

Quick Start

Using the CLI

Run a complete analysis with default synthetic data:

python main.py

Run with custom configuration:

python main.py --config config/custom_config.yaml --output results/

Analyze stock data from Yahoo Finance:

python main.py --data-source yfinance --symbol AAPL --start-date 2020-01-01 --end-date 2023-01-01

Using the Web Interface

Launch the Streamlit application:

streamlit run streamlit_app.py

Then open your browser to http://localhost:8501 to access the interactive interface.

Using the Python API

from src.trend_analyzer import TrendAnalyzer
from src.visualizer import TimeSeriesVisualizer
from src.forecaster import TimeSeriesForecaster

# Initialize components
analyzer = TrendAnalyzer()
visualizer = TimeSeriesVisualizer()
forecaster = TimeSeriesForecaster()

# Generate synthetic data
data = analyzer.generate_synthetic_data()

# Perform trend analysis
ma_trend = analyzer.moving_average_trend()
poly_trend = analyzer.polynomial_trend()

# Detect anomalies
anomaly_scores, anomaly_labels = analyzer.detect_anomalies()

# Generate forecasts
forecast = forecaster.linear_regression_forecast(data, forecast_horizon=30)

# Create visualizations
visualizer.plot_original_vs_trends(data, analyzer.trend_components)

Configuration

The project uses YAML configuration files for customization. See config/config.yaml for available options:

data:
  source: "synthetic"  # Options: "synthetic", "yfinance", "file"
  synthetic:
    length: 1000
    trend_type: "linear"  # Options: "linear", "quadratic", "exponential"
    seasonality: true
    noise_level: 0.1

models:
  moving_average:
    window: 30
  polynomial:
    degree: 2
  arima:
    order: (1, 1, 1)
  prophet:
    yearly_seasonality: true
    weekly_seasonality: true

visualization:
  figure_size: [12, 8]
  save_plots: true

Project Structure

time-series-trend-analysis/
├── src/                    # Source code
│   ├── trend_analyzer.py   # Core trend analysis functionality
│   ├── visualizer.py      # Visualization components
│   └── forecaster.py      # Forecasting methods
├── config/                 # Configuration files
│   └── config.yaml        # Main configuration
├── tests/                  # Unit tests
│   └── test_components.py # Test cases
├── notebooks/              # Jupyter notebooks (optional)
├── data/                   # Data directory
├── results/                # Output results
├── plots/                  # Generated plots
├── models/                 # Saved models
├── main.py                 # CLI entry point
├── streamlit_app.py        # Web interface
├── requirements.txt        # Python dependencies
├── .gitignore             # Git ignore rules
└── README.md              # This file

Methods and Algorithms

Trend Analysis

• Moving Average: Smoothing using rolling window averages
• Polynomial Regression: Fitting polynomial trends of various degrees
• Detrending: Removing trend components for further analysis

Anomaly Detection

• Isolation Forest: Unsupervised anomaly detection using isolation trees
• Statistical Methods: Z-score based anomaly detection

Forecasting

• ARIMA: AutoRegressive Integrated Moving Average models
• Prophet: Facebook's forecasting tool for business time series
• LSTM: Long Short-Term Memory neural networks
• Linear Regression: Feature-engineered linear models

Visualization

• Trend Comparison: Original data vs extracted trends
• Detrending Analysis: Before/after detrending visualization
• Anomaly Detection: Highlighted anomalies with scores
• Forecast Comparison: Multiple forecasting methods side-by-side
• Comprehensive Dashboard: All analyses in one view

Testing

Run the test suite:

python -m pytest tests/ -v

Or run specific test modules:

python tests/test_components.py

Examples

Basic Trend Analysis

from src.trend_analyzer import TrendAnalyzer

analyzer = TrendAnalyzer()
data = analyzer.generate_synthetic_data()

# Extract trends
ma_trend = analyzer.moving_average_trend(window=20)
poly_trend = analyzer.polynomial_trend(degree=2)

# Detrend the data
detrended = analyzer.detrend("polynomial")

Anomaly Detection

# Detect anomalies using Isolation Forest
anomaly_scores, anomaly_labels = analyzer.detect_anomalies("isolation_forest")

# Detect anomalies using statistical method
anomaly_scores, anomaly_labels = analyzer.detect_anomalies("statistical")

Forecasting

from src.forecaster import TimeSeriesForecaster

forecaster = TimeSeriesForecaster()

# Generate forecasts using multiple methods
arima_forecast = forecaster.arima_forecast(data, forecast_horizon=30)
prophet_forecast = forecaster.prophet_forecast(data, forecast_horizon=30)
lstm_forecast = forecaster.lstm_forecast(data, forecast_horizon=30)

# Create ensemble forecast
forecasts = {
    'ARIMA': arima_forecast,
    'Prophet': prophet_forecast,
    'LSTM': lstm_forecast
}
ensemble = forecaster.ensemble_forecast(forecasts)

Visualization

from src.visualizer import TimeSeriesVisualizer

visualizer = TimeSeriesVisualizer()

# Plot trend analysis
visualizer.plot_original_vs_trends(data, analyzer.trend_components)

# Plot anomaly detection
visualizer.plot_anomaly_detection(data, anomaly_scores, anomaly_labels)

# Create comprehensive dashboard
visualizer.create_dashboard(
    data, 
    analyzer.trend_components, 
    detrended, 
    anomaly_scores, 
    anomaly_labels
)

Command Line Options

python main.py --help

Available options:

• --config: Path to configuration YAML file
• --output: Output directory for results and plots
• --data-source: Data source (synthetic, yfinance, file)
• --symbol: Stock symbol for yfinance data
• --start-date: Start date for yfinance data (YYYY-MM-DD)
• --end-date: End date for yfinance data (YYYY-MM-DD)
• --verbose: Enable verbose logging

Dependencies

Core Dependencies

• numpy >= 1.24.0
• pandas >= 2.0.0
• scikit-learn >= 1.3.0
• matplotlib >= 3.7.0
• seaborn >= 0.12.0

Optional Dependencies

• statsmodels >= 0.14.0 (for ARIMA)
• prophet >= 1.1.0 (for Prophet forecasting)
• torch >= 2.0.0 (for LSTM)
• yfinance >= 0.2.0 (for stock data)
• streamlit >= 1.25.0 (for web interface)

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/new-feature)
3. Commit your changes (git commit -am 'Add new feature')
4. Push to the branch (git push origin feature/new-feature)
5. Create a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Facebook Prophet team for the excellent forecasting library
• Statsmodels contributors for comprehensive statistical tools
• PyTorch team for deep learning capabilities
• Streamlit team for the web interface framework

Support

For questions, issues, or contributions, please:

1. Check the existing issues on GitHub
2. Create a new issue with detailed description
3. Contact the maintainers

Changelog

Version 1.0.0

• Initial release with comprehensive trend analysis
• Support for multiple forecasting methods
• Interactive web interface
• CLI tool for batch processing
• Comprehensive test suite
• YAML configuration system

Time-Series-Trend-Analysis