Future Proofing Spirits

Forecasting Alcohol Consumption Trends for Generations Alpha & Beta

Project Overview

This project builds an automated ETL-P (Extract, Transform, Load, Predict) pipeline to ingest historical global alcohol consumption data and forecast future demand patterns. Specifically, it targets Generation Alpha (2010–2024) and Generation Beta (2025–2039) as they enter legal drinking age, helping stakeholders pivot strategies for R&D, supply chain, and healthcare planning.

Core Objectives:

1. Model Historical Trends: Analyze consumption patterns of Gen X, Millennials, and Gen Z.
2. Forecast Demand: Project "entry-level" consumption volume (liters of pure alcohol) for Gen Alpha/Beta for the period 2025–2045.
3. Identify Shifts: Detect market contractions ("sober curious" movement) or expansions in specific regions.

---

Architecture

The system follows a modular architecture:

1. Ingestion: Scrapes/Downloads data from Our World in Data (OWID) & WHO.
2. Processing: Cleans data, imputes missing values, and maps birth years to generational cohorts.
3. Storage: Centralized Data Warehouse (PostgreSQL/Snowflake).
4. Machine Learning: ARIMA/Prophet models for time-series forecasting.
5. Visualization: Interactive dashboards for strategic insights.

---

📂 Project Structure

alcohol-consumption-prediction/
├── data/
│   ├── raw/                   # Landed raw data (CSV/JSON)
│   ├── processed/             # Cleaned & standardized data
│   └── external/              # Auxiliary data (Population/Demographics)
├── src/
│   ├── etl/                   # Extract, Transform, Load scripts
│   ├── models/                # ML Training & Prediction scripts
│   └── visualization/         # Dashboard code
├── notebooks/                 # Jupyter notebooks for EDA
├── storage/                   # Technical documentation & Specs
├── requirements.txt           # Python dependencies
└── README.md                  # You are here

---

Getting Started

Prerequisites

• Python 3.9+
• Docker & Docker Compose (for local DB)

Installation

1. Clone the repository

   git clone https://github.com/your-username/alcohol-consumption-genz.git
   cd alcohol-consumption-genz

2. Set up Virtual Environment

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

3. Launch Infrastructure (Local)

   docker-compose up -d

   • This starts PostgreSQL.

Usage

Running the ETL Pipeline:

# Manual run via CLI (for dev)
python src/etl/extract.py
python src/etl/transform.py
python src/etl/load.py

Training Models:

python src/models/training.py --model prophet --epochs 100

Generating Forecasts:

python src/models/predict.py --start-year 2025 --end-year 2045

---

🗺️ Roadmap

We follow a phased release strategy. See the full roadmap in storage/tickets.md.

• Planning: PRD & Architecture Design
• v0.1: Infrastructure & Data Ingestion
• v0.2: Data Warehouse & Processing
• v0.3: ML Forecasting Models
• v1.0: Visualization Dashboard

---

Documentation

• Product Requirements Document (PRD)
• Technical Guide
• Architecture Diagram

---

License

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