Project Dionysus

Project Dionysus is an educational time series analysis project focused on modeling and forecasting wine production data from multiple producers.

Overview

The project demonstrates how to:

• Load and preprocess real-world time series datasets with inconsistent date formats
• Structure reusable, scalable analysis code via object-oriented design
• Generate summary statistics and forecasts using SARIMAX modeling

Data

Three sample datasets are provided:

• Almeirim.csv
• Benavente.csv
• Cartaxo.csv

Each contains monthly production estimates (in thousands of liters) over a shared time interval. The datasets reflect three distinct formatting styles, representative of future scalability challenges.

Key Components

• Wine Class: Encapsulates data loading, validation, reporting, and forecasting logic
• report() Method: Outputs filtered summary statistics over a user-defined date range
• forecast() Method: Uses SARIMAX to predict future output and visualize trends
• Automation: Automatically processes all CSV files in the /data/ directory

Requirements

Install dependencies:

pip install -r requirements.txt