Argentine Fintech Screener 🇦🇷

Hi! Thanks for stopping by my portfolio. I'm currently a student learning Data Engineering, and I built this project to practice processing real financial data and building high-performance interactive apps from scratch.

This project analyzes the extreme volatility of the Argentine market (MERVAL index, ADRs, and implicit FX rates) using the Modern Data Stack 2026.

🚀 Why the Argentine Market?

When learning Time Series analysis and Anomaly Detection, you need data that actually moves. I chose the Argentine market because its extreme volatility, high inflation, and sudden political shocks make it a world-class laboratory for data science.

Instead of looking at a flat chart, tracking the Argentine market during election cycles shows you exactly what a massive economic swing looks like in the data.

🔬 Digital Signal Processing (DSP) Implementation

I implemented a Savitzky-Golay low-pass filter to calculate a Signal-to-Noise Ratio (SNR). This allows the tool to mathematically distinguish whether an asset is moving due to underlying Market Trends (the core signal) or purely due to short-term Political Noise and panic.

🛠️ The 2026 Modern Data Stack

While I have mastery over traditional tools like Pandas and SQLite, I deliberately chose these technologies to push the boundaries of performance and developer experience:

• Polars: Used for lightning-fast data pipelines and vectorized operations. I optimized event matching using join_asof for high-performance temporal joins.
• DuckDB: Serving as the local analytical engine, allowing complex SQL queries without the overhead of a traditional server.
• Marimo: A reactive notebook system used to build an interactive UI that behaves like a modern web app.

🏗️ Project Architecture

graph LR
    A[Yahoo Finance API] --> B[fetch_data.py]
    B --> C{src/data_ingestion}
    C --> D[(DuckDB: merval_data.db)]
    D --> E[dashboard.py]
    E --> F{src/processing}
    F --> G[Interactive UI / Plots]
    F --> H[Static HTML Export]

Loading

📸 Showcase & Demos

Interactive Dashboard

Note: This dashboard allows filtering by political events and visualizing denoised trends.

Market SNR Table

Assets with lower SNR ratios indicate higher sensitivity to political "noise".

📄 Viewing Without Running

If you just want to see the results without installing anything, check the Live Interactive Dashboard or browse the screenshots in the docs/assets/ folder.

🚦 Getting Started

Prerequisites

• Python 3.10 or higher.

1. Environment Setup

# Clone the repository
git clone https://github.com/your-username/argentinian-markets.git
cd argentinian-markets

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

# Install dependencies
pip install -r requirements.txt

2. Data Ingestion

Populate the local DuckDB database with the latest 3 years of market data:

python fetch_data.py

3. Run the Dashboard

Launch the interactive Marimo server:

python -m marimo edit dashboard.py

---

This project is part of my path towards becoming a Senior Data Engineer. Feel free to reach out for feedback or collaboration!