Timeseries Compute

Overview

████████╗██╗███╗   ███╗███████╗███████╗███████╗██████╗ ██╗███████╗███████╗
╚══██╔══╝██║████╗ ████║██╔════╝██╔════╝██╔════╝██╔══██╗██║██╔════╝██╔════╝
   ██║   ██║██╔████╔██║█████╗  ███████╗█████╗  ██████╔╝██║█████╗g ███████╗
   ██║   ██║██║╚██╔╝██║██╔══╝  ╚════██║██╔══╝  ██╔══██╗██║██╔══╝m ╚════██║
   ██║   ██║██║ ╚═╝ ██║███████╗███████║███████╗██║  ██║██║███████╗███████║
   ╚═╝   ╚═╝╚═╝     ╚═╝╚══════╝╚══════╝╚══════╝╚═╝  ╚═╝╚═╝╚══════╝╚══════╝
             ██████╗ ██████╗ ███╗   ███╗██████╗ ██╗   ██╗████████╗███████╗
            ██╔════╝██╔═══██╗████╗ ████║██╔══██╗██║   ██║╚══██╔══╝██╔════╝
            ██║     ██║   ██║██╔████╔██║██████╔╝██║   ██║   ██║   █████╗
            ██║     ██║   ██║██║╚██╔╝██║██╔═══╝ ██║   ██║   ██║   ██╔══╝
            ╚██████╗╚██████╔╝██║ ╚═╝ ██║██║     ╚██████╔╝   ██║   ███████╗
             ╚═════╝ ╚═════╝ ╚═╝     ╚═╝╚═╝      ╚═════╝    ╚═╝   ╚══════╝

Implementation hosted at www.spilloverlab.com.

A Python package for timeseries data processing and modeling using ARIMA and GARCH models with both univariate and multivariate capabilities.

Features

• Price series generation for single and multiple assets
• Data preprocessing with configurable missing data handling and scaling options
• Stationarity testing and transformation for time series analysis
• ARIMA modeling for time series forecasting
• GARCH modeling for volatility forecasting and risk assessment
• Bivariate GARCH modeling with both Constant Conditional Correlation (CCC) and Dynamic Conditional Correlation (DCC) methods
• EWMA covariance calculation for dynamic correlation analysis
• Portfolio risk assessment using volatility and correlation matrices
• Market spillover effects analysis with Granger causality testing and shock transmission modeling
• Visualization tools for interpreting complex market interactions and spillover relationships

Integration Overview

flowchart TB
    %% Styling
    classDef person fill:#7B4B94,color:#fff,stroke:#5D2B6D,stroke-width:1px
    classDef agent fill:#7B4B94,color:#fff,stroke:#5D2B6D,stroke-width:1px
    classDef system fill:#1168BD,color:#fff,stroke:#0B4884,stroke-width:1px
    classDef external fill:#999999,color:#fff,stroke:#6B6B6B,stroke-width:1px
    classDef database fill:#2E7C8F,color:#fff,stroke:#1D4E5E,stroke-width:1px
    classDef publishing fill:#E67E22,color:#fff,stroke:#D35400,stroke-width:1px
    
    %% Actors and Systems
    User((User)):::person
    AIAgent((AI Agent)):::agent
    
    %% Main Systems
    TimeSeriesFrontend["Frontend App"]:::system
    TimeSeriesPipeline["RESTful Pipeline"]:::system
    MCPServer["MCP Server"]:::system
    TimeseriesCompute["Timeseries-Compute 
    Python Package"]:::system
    
    %% Database
    TimeSeriesDB[("Relational database")]:::database
    
    %% External Systems
    ExternalDataSource[(Yahoo Finance / Stooq)]:::external
    
    %% Publishing Platforms
    PublishingPlatforms["
    GitHub
    Docker Hub
    Google Cloud Run
    PyPI
    Read the Docs"]:::publishing
    
    %% Relationships
    User -- "Uses UI" --> TimeSeriesFrontend
    AIAgent -- "Natural language requests" --> MCPServer
    TimeSeriesFrontend -- "Makes API calls to" --> TimeSeriesPipeline
    MCPServer -- "Makes API calls to" --> TimeSeriesPipeline
    TimeSeriesPipeline -- "Inserts results into" --> TimeSeriesDB
    TimeSeriesPipeline -- "imports" --> TimeseriesCompute
    User -- "pip install" --> TimeseriesCompute
    AIAgent -- "pip install" --> TimeseriesCompute
    ExternalDataSource -- "Provides time series data" --> TimeSeriesPipeline
    
    %% Publishing relationships (simplified)
    TimeSeriesFrontend  --> PublishingPlatforms
    TimeSeriesPipeline --> PublishingPlatforms
    TimeseriesCompute --> PublishingPlatforms

Quick Start

Installation

Using uv (fastest):

# Install uv
pip install uv
# create venv
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt

Using venv (classic):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install timeseries-compute

Install from GitHub using venv (latest development version):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install git+https://github.com/garthmortensen/timeseries-compute.git

Example Usage

For univariate time series analysis:

python -m timeseries_compute.examples.example_univariate_garch

For multivariate GARCH analysis (correlation between two assets):

python -m timeseries_compute.examples.example_multivariate_garch

Docker Support

Run with Docker for isolated environments:

# build the image
docker build -t timeseries-compute:latest ./

# Run the univariate example
docker run -it timeseries-compute:latest /app/timeseries_compute/examples/example_univariate_garch.py

# Run the multivariate example
docker run -it timeseries-compute:latest /app/timeseries_compute/examples/example_multivariate_garch.py

# Get into interactive shell
docker run -it --entrypoint /bin/bash timeseries-compute:latest

Project Structure

timeseries_compute/......................
├── __init__.py                         # Package initialization and public API
├── data_generator.py                   # Synthetic price data generation with random walks and statistical properties
├── data_processor.py                   # Data transformation, missing value handling, scaling, and stationarity testing
├── export_util.py                      # Data export utilities for tracking analysis lineage
├── spillover_processor.py              # Diebold-Yilmaz spillover analysis and Granger causality testing
├── stats_model.py                      # ARIMA, GARCH, and multivariate GARCH model implementations
├── examples/............................
│   ├── __init__.py                     # Makes examples importable as a module
│   ├── example_multivariate_garch.py   # Correlation analysis between multiple markets with CC-GARCH and DCC-GARCH
│   └── example_univariate_garch.py     # Basic ARIMA and GARCH modeling for single-series forecasting
└── tests/...............................
    ├── __init__.py                     # Makes tests discoverable by pytest
    ├── test_data_generator_advanced.py # Advanced data generation features and statistical property testing
    ├── test_data_generator.py          # Basic price generation functionality testing
    ├── test_data_processor.py          # Data transformation, scaling, and stationarity testing
    ├── test_spillover_processor.py     # Spillover analysis and Granger causality testing
    ├── test_stats_model_arima.py       # ARIMA modeling with specialized fixtures and edge cases
    └── test_stats_model_garch.py       # GARCH volatility modeling with different distributions

Architectural Diagrams

Level 2: Container Diagram

flowchart TB
    %% Styling
    classDef person fill:#08427B,color:#fff,stroke:#052E56,stroke-width:1px
    classDef container fill:#438DD5,color:#fff,stroke:#2E6295,stroke-width:1px
    classDef external fill:#999999,color:#fff,stroke:#6B6B6B,stroke-width:1px
    classDef system fill:#1168BD,color:#fff,stroke:#0B4884,stroke-width:1px
    
    %% Person
    User((User)):::person
    
    %% System boundary
    subgraph TimeseriesComputeSystem["Timeseries Compute System"]
        PythonPackage["Python Package<br>[Library]<br>Core functions for analysis"]:::container
        Dockerized["Docker Container<br>[Linux]<br>Containerized deployment"]:::container
        ExampleScripts["Example Scripts<br>[Python]<br>Demonstration use cases"]:::container
        TestSuite["Test Suite<br>[pytest]<br>Validates package functionality"]:::container
        CIpipeline["CI/CD Pipeline<br>[GitHub Actions]<br>Automates testing/deployment"]:::container
        Documentation["Documentation<br>[ReadTheDocs]<br>API and usage docs"]:::container
    end
    
    %% External Systems
    ExternalDataSource[(External Data Source)]:::external
    AnalysisTool[Analysis & Visualization Tools]:::external
    PyPI[PyPI Repository]:::external
    DockerHub[Docker Hub Repository]:::external
    
    %% Relationships
    User -- "Imports [Python]" --> PythonPackage
    User -- "Runs [CLI]" --> ExampleScripts
    User -- "Reads [Web]" --> Documentation
    ExampleScripts -- "Uses" --> PythonPackage
    TestSuite -- "Tests" --> PythonPackage
    PythonPackage -- "Packaged into" --> Dockerized
    CIpipeline -- "Builds and tests" --> Dockerized
    CIpipeline -- "Runs" --> TestSuite
    CIpipeline -- "Publishes" --> PyPI
    CIpipeline -- "Publishes" --> DockerHub
    CIpipeline -- "Updates" --> Documentation
    ExternalDataSource -- "Provides data to" --> PythonPackage
    PythonPackage -- "Exports analysis to" --> AnalysisTool
    User -- "Downloads from" --> PyPI
    User -- "Runs with" --> DockerHub

Level 3: Component Diagram

flowchart TB
    %% Styling
    classDef person fill:#08427B,color:#fff,stroke:#052E56,stroke-width:1px
    classDef component fill:#85BBF0,color:#000,stroke:#5D82A8,stroke-width:1px
    classDef container fill:#438DD5,color:#fff,stroke:#2E6295,stroke-width:1px
    classDef external fill:#999999,color:#fff,stroke:#6B6B6B,stroke-width:1px
    
    %% Person
    User((User)):::person
    
    %% Package Container
    subgraph PythonPackage["Python Package"]
        DataGenerator["Data Generator<br>[Python]<br>Creates synthetic time series"]:::component
        DataProcessor["Data Processor<br>[Python]<br>Transforms and tests data"]:::component
        StatsModels["Statistical Models<br>[Python]<br>ARIMA and GARCH models"]:::component
        SpilloverProcessor["Spillover Processor<br>[Python]<br>Market interaction analysis"]:::component
        ExportUtil["Export Utility<br>[Python]<br>Data export functions"]:::component
        ExampleScripts["Example Scripts<br>[Python]<br>Usage demonstrations"]:::component
        TestSuite["Test Suite<br>[pytest]<br>Validates functionality"]:::component
        
        %% Component relationships
        ExampleScripts --> DataGenerator
        ExampleScripts --> DataProcessor
        ExampleScripts --> StatsModels
        ExampleScripts --> SpilloverProcessor
        ExampleScripts --> ExportUtil
        StatsModels --> DataProcessor
        SpilloverProcessor --> StatsModels
        SpilloverProcessor --> DataProcessor
        TestSuite --> DataGenerator
        TestSuite --> DataProcessor
        TestSuite --> StatsModels
        TestSuite --> SpilloverProcessor
        TestSuite --> ExportUtil
    end

    %% External Components
    StatsLibraries[(Statistical Libraries<br>statsmodels, arch)]:::external
    DataLibraries[(Data Libraries<br>pandas, numpy)]:::external
    VisualizationLibraries[(Visualization<br>matplotlib)]:::external
    
    %% Relationships
    User -- "Uses" --> ExampleScripts
    User -- "Uses directly" --> DataGenerator
    User -- "Uses directly" --> DataProcessor
    User -- "Uses directly" --> StatsModels
    User -- "Uses directly" --> SpilloverProcessor
    DataGenerator -- "Uses" --> DataLibraries
    DataProcessor -- "Uses" --> DataLibraries
    StatsModels -- "Uses" --> StatsLibraries
    StatsModels -- "Uses" --> DataLibraries
    ExampleScripts -- "Uses" --> VisualizationLibraries
    SpilloverProcessor -- "Uses" --> VisualizationLibraries
    ExportUtil -- "Uses" --> DataLibraries

Level 4: Code/Class Diagram

classDiagram
    %% Main Classes (actual)
    class PriceSeriesGenerator {
        +start_date: str
        +end_date: str
        +dates: pd.DatetimeIndex
        +__init__(start_date, end_date)
        +generate_correlated_prices(anchor_prices, correlation_matrix): Dict[str, list]
    }
    
    class MissingDataHandler {
        +__init__()
        +drop_na(data): pd.DataFrame
        +forward_fill(data): pd.DataFrame
    }
    
    class DataScaler {
        +scale_data_standardize(data): pd.DataFrame
        +scale_data_minmax(data): pd.DataFrame
    }
    
    class StationaryReturnsProcessor {
        +make_stationary(data, method): pd.DataFrame
        +test_stationarity(data, test): Dict
        +log_adf_results(data, p_value_threshold): None
    }
    
    class ModelARIMA {
        +data: pd.DataFrame
        +order: Tuple[int, int, int]
        +steps: int
        +models: Dict[str, ARIMA]
        +fits: Dict[str, ARIMA]
        +__init__(data, order, steps)
        +fit(): Dict[str, ARIMA]
        +summary(): Dict[str, str]
        +forecast(): Dict[str, Union[float, list]]
    }
    
    class ModelGARCH {
        +data: pd.DataFrame
        +p: int
        +q: int
        +dist: str
        +models: Dict[str, arch_model]
        +fits: Dict[str, arch_model]
        +__init__(data, p, q, dist)
        +fit(): Dict[str, arch_model]
        +summary(): Dict[str, str]
        +forecast(steps): Dict[str, float]
    }
    
    class ModelMultivariateGARCH {
        +data: pd.DataFrame
        +p: int
        +q: int
        +model_type: str
        +fits: Dict
        +cc_results: Dict
        +dcc_results: Dict
        +__init__(data, p, q, model_type)
        +fit_cc_garch(): Dict[str, Any]
        +fit_dcc_garch(lambda_val): Dict[str, Any]
    }
    
    %% Factory Classes
    class ModelFactory {
        <<static>>
        +create_model(model_type, data, order, steps, p, q, dist, mv_model_type): Union[ModelARIMA, ModelGARCH, ModelMultivariateGARCH]
    }
    
    class MissingDataHandlerFactory {
        <<static>>
        +create_handler(strategy): Callable[[pd.DataFrame], pd.DataFrame]
    }
    
    class DataScalerFactory {
        <<static>>
        +create_handler(strategy): Callable[[pd.DataFrame], pd.DataFrame]
    }
    
    class StationaryReturnsProcessorFactory {
        <<static>>
        +create_handler(strategy): StationaryReturnsProcessor
    }
    
    %% Module-level Functions (actual implementation structure)
    class DataGeneratorModule {
        <<module>>
        +set_random_seed(seed): None
        +generate_price_series(start_date, end_date, anchor_prices, random_seed, correlations): Tuple[Dict, pd.DataFrame]
    }
    
    class DataProcessorModule {
        <<module>>
        +fill_data(df, strategy): pd.DataFrame
        +scale_data(df, method): pd.DataFrame
        +scale_for_garch(df, target_scale): pd.DataFrame
        +stationarize_data(df, method): pd.DataFrame
        +test_stationarity(df, method): Dict
        +log_stationarity(adf_results, p_value_threshold): None
        +price_to_returns(prices): pd.DataFrame
        +prepare_timeseries_data(df): pd.DataFrame
        +calculate_ewma_covariance(series1, series2, lambda_val): pd.Series
        +calculate_ewma_volatility(series, lambda_val): pd.Series
    }
    
    class StatsModelModule {
        <<module>>
        +run_arima(df_stationary, p, d, q, forecast_steps): Tuple[Dict, Dict]
        +run_garch(df_stationary, p, q, dist, forecast_steps): Tuple[Dict, Dict]
        +run_multivariate_garch(df_stationary, arima_fits, garch_fits, lambda_val): Dict
        +calculate_correlation_matrix(standardized_residuals): pd.DataFrame
        +calculate_dynamic_correlation(ewma_cov, ewma_vol1, ewma_vol2): pd.Series
        +construct_covariance_matrix(volatilities, correlation): np.ndarray
        +calculate_portfolio_risk(weights, cov_matrix): Tuple[float, float]
        +calculate_stats(series): Dict
    }

    class SpilloverProcessorModule {
        <<module>>
        +test_granger_causality(series1, series2, max_lag, significance_level): Dict
        +analyze_shock_spillover(residuals1, volatility2, max_lag): Dict
        +run_spillover_analysis(df_stationary, arima_fits, garch_fits, lambda_val, max_lag, significance_level): Dict
    }
    
    class ExportUtilModule {
        <<module>>
        +export_data(data, folder, name): Any
    }
    
    %% Example Scripts
    class ExampleUnivariateGARCH {
        <<script>>
        +main(): None
    }
    
    class ExampleMultivariateGARCH {
        <<script>>
        +main(): None
    }
    
    %% Relationships - Factory patterns
    MissingDataHandlerFactory --> MissingDataHandler: creates
    DataScalerFactory --> DataScaler: creates
    StationaryReturnsProcessorFactory --> StationaryReturnsProcessor: creates
    ModelFactory --> ModelARIMA: creates
    ModelFactory --> ModelGARCH: creates
    ModelFactory --> ModelMultivariateGARCH: creates
    
    %% Module dependencies
    DataProcessorModule --> MissingDataHandler: uses
    DataProcessorModule --> DataScaler: uses
    DataProcessorModule --> StationaryReturnsProcessor: uses
    DataProcessorModule --> MissingDataHandlerFactory: uses
    DataProcessorModule --> DataScalerFactory: uses
    DataProcessorModule --> StationaryReturnsProcessorFactory: uses
    
    StatsModelModule --> ModelARIMA: uses
    StatsModelModule --> ModelGARCH: uses
    StatsModelModule --> ModelMultivariateGARCH: uses
    StatsModelModule --> ModelFactory: uses
    StatsModelModule --> DataProcessorModule: uses
    
    SpilloverProcessorModule --> StatsModelModule: uses
    SpilloverProcessorModule --> DataProcessorModule: uses
    
    %% Example script dependencies
    ExampleUnivariateGARCH --> DataGeneratorModule: uses
    ExampleUnivariateGARCH --> DataProcessorModule: uses
    ExampleUnivariateGARCH --> StatsModelModule: uses
    
    ExampleMultivariateGARCH --> DataGeneratorModule: uses
    ExampleMultivariateGARCH --> DataProcessorModule: uses
    ExampleMultivariateGARCH --> StatsModelModule: uses
    ExampleMultivariateGARCH --> ExportUtilModule: uses
    
    %% Core class usage
    DataGeneratorModule --> PriceSeriesGenerator: uses

CI/CD Process

• Triggers: Runs when code is pushed to branches main or dev
• pytest: Validates code across multiple Python versions and OS
• Building: Creates package distributions and documentation
• Publishing: Deploys to PyPI, Docker Hub and ReadTheDocs

flowchart TB
    %% Styling
    classDef person fill:#08427B,color:#fff,stroke:#052E56,stroke-width:1px
    classDef system fill:#1168BD,color:#fff,stroke:#0B4884,stroke-width:1px
    classDef external fill:#999999,color:#fff,stroke:#6B6B6B,stroke-width:1px
    classDef pipeline fill:#ff9900,color:#fff,stroke:#cc7700,stroke-width:1px
    
    %% Actors
    Developer((Developer)):::person
    
    %% Main Systems
    TimeseriesCompute["Timeseries Compute\nPython Package"]:::system
    
    %% Source Control
    GitHub["GitHub\nSource Repository"]:::external
    
    %% CI/CD Pipeline and Tools
    GitHubActions["GitHub Actions\nCI/CD Pipeline"]:::pipeline
    
    %% Distribution Platforms
    PyPI["PyPI Registry"]:::external
    DockerHub["Docker Hub"]:::external
    ReadTheDocs["ReadTheDocs"]:::external
    
    %% Code Quality Services
    Codecov["Codecov\nCode Coverage"]:::external
    
    %% Flow
    Developer -- "Commits code to" --> GitHub
    GitHub -- "Triggers on push\nto main/dev" --> GitHubActions
    
    %% Primary Jobs
    subgraph TestJob["Test Job"]
        Test["Run Tests\nPytest"]:::pipeline
        Lint["Lint with Flake8"]:::pipeline
        
        Lint --> Test
    end
    
    subgraph DockerJob["Docker Job"]
        BuildDocker["Build Docker Image"]:::pipeline
    end
    
    subgraph BuildJob["Build Job"]
        BuildPackage["Build Package\nSDist & Wheel"]:::pipeline
        VerifyPackage["Verify with Twine"]:::pipeline
        
        BuildPackage --> VerifyPackage
    end
    
    subgraph DocsJob["Docs Job"]
        BuildDocs["Generate Docs\nSphinx"]:::pipeline
        BuildUML["Generate UML\nDiagrams"]:::pipeline
        
        BuildDocs --> BuildUML
    end
    
    subgraph PublishJob["Publish Job"]
        PublishPyPI["Publish to PyPI"]:::pipeline
    end
    
    %% Job Dependencies
    GitHubActions --> TestJob
    
    TestJob --> DockerJob
    TestJob --> BuildJob
    TestJob --> DocsJob
    
    BuildJob --> PublishJob
    DocsJob --> PublishJob
    
    %% External Services Connections
    Test -- "Upload Results" --> Codecov
    BuildDocker -- "Push Image" --> DockerHub
    DocsJob -- "Deploy Documentation" --> ReadTheDocs
    PublishPyPI -- "Deploy Package" --> PyPI
    
    %% Final Products
    PyPI --> TimeseriesCompute
    DockerHub --> TimeseriesCompute
    ReadTheDocs -- "Documents" --> TimeseriesCompute

Development

Environment Setup

Option 1 (recommended):

mkdir timeseries-compute
cd timeseries-compute

# create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

pip install timeseries-compute

Option 2:

# clone the repository
git clone https://github.com/garthmortensen/timeseries-compute.git
cd timeseries-compute

# create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

pip install -e ".[dev]"

Testing

pytest --cov=timeseries_compute

Tag & Publish

Bump version in pyproject.toml and README.md

git add pyproject.toml README.md
git commit -m "version bump"
git tag v0.2.41
git push && git push --tags

Methodology

This section illustrates the statistical workflows and logic for the key implementations in the package.

Spillover Analysis (Diebold-Yilmaz)

graph TD
    A[Multivariate Returns] --> B[Fit VAR Model]
    B --> C[Calculate FEVD Matrix]
    C --> D[Extract Spillover Indices]
    D --> E[Total Connectedness Index]
    D --> F[Directional Spillovers]
    D --> G[Net Spillovers]
    
    style A fill:#e1f5fe
    style E fill:#c8e6c9
    style F fill:#c8e6c9
    style G fill:#c8e6c9

ARIMA Modeling Logic

graph TD
    A[Stationary Returns] --> B["Fit ARIMA(p,d,q)"]
    B --> C[Extract Residuals]
    B --> D[Generate Forecasts]
    C --> E[Filter Conditional Mean]
    
    style A fill:#e1f5fe
    style D fill:#c8e6c9
    style E fill:#fff3e0

GARCH Modeling Logic

graph TD
    A[Stationary Returns] --> B["Fit GARCH(p,q)"]
    B --> C[Extract Conditional Volatility]
    B --> D[Generate Volatility Forecasts]
    C --> E[Calculate Standardized Residuals]
    
    style A fill:#e1f5fe
    style D fill:#c8e6c9
    style C fill:#f3e5f5
    style E fill:#e8f5e8

Multivariate GARCH Logic

graph TD
    A[Multiple Return Series] --> B[Filter Conditional Means]
    B --> C[Model Individual Volatilities]
    C --> D[Extract Standardized Residuals]
    
    D --> E[Constant Correlation]
    D --> F[Dynamic Correlation]
    
    E --> G[CCC Covariance Matrix]
    F --> H[DCC Correlation Series]
    
    style A fill:#e1f5fe
    style G fill:#c8e6c9
    style H fill:#c8e6c9

Complete Statistical Workflow

graph TD
    A[Financial Data] --> B[Data Preprocessing]
    B --> C[Stationarity Testing]
    C --> D[Mean Modeling]
    D --> E[Volatility Modeling]
    E --> F[Correlation Analysis]
    F --> G[Risk Assessment]
    
    style A fill:#ffebee
    style G fill:#c8e6c9