Federico Ariton

Olist E-commerce Marketing Analytics

A full-stack marketing analytics pipeline simulating a real-world e-commerce business case. Using Brazil’s Olist marketplace dataset, I built a PostgreSQL database, engineered KPIs, and performed product segmentation using PCA + KMeans. The result is a Streamlit dashboard that enables strategic decisions for marketing and inventory optimization.

Skills: PostgreSQL, Docker, ETL (Pandas), PCA, KMeans, Streamlit, Plotly

Stock Price Forecasting Using Sentiment and Big Data

This project forecasts 1-, 3-, and 7-day stock prices by combining sentiment analysis from Twitter with market data. It uses statistical models (ARIMAX, SARIMAX) and deep learning models (LSTM, GRU), all powered by a scalable big data architecture.

Built on a Lambda architecture with Apache Spark and HDFS, it benchmarks four databases and delivers real-time predictions via a Streamlit dashboard.

Skills: PySpark, HDFS, Spark NLP, ARIMAX, LSTM, GRU, YCSB, Dashboarding

AI Super-Resolution in Games: Real-ESRGAN vs. SRCNN

This Master’s project compares two AI models SRCNN and Real-ESRGAN for enhancing retro game visuals by reconstructing pixel details. These models upscale low-res images while reducing storage needs producing images up to 10× smaller in file size.

A big data pipeline using Hadoop HDFS and Apache Spark enabled scalable processing of game screenshots and high-resolution training data (DIV2K).

Skills: PyTorch, HDFS, Spark, OpenCV, Image Upscaling, PSNR/SSIM, Neural Networks

Predicting and Forecasting Agricultural Export Values in Ireland

This project focuses on analyzing production-related factors that influence agricultural export values in Ireland. It employs statistical and predictive modeling techniques to uncover trends and provides actionable insights for optimizing production and pricing strategies.

Skills: Data analysis, predictive modeling, dashboard development, CRISP-DM methodology

Customer Segmentation for E-commerce

This project applies clustering techniques to an e-commerce dataset to segment customers based on their behavior. Customer segmentation enables businesses to tailor marketing strategies, improve customer engagement, and increase overall profitability.

Skills: Customer Segmentation, Clustering

Database Design and Queries with MySQL

Designed and normalized a relational database schema for employee management, achieving 3NF. Wrote advanced SQL queries for data retrieval, aggregation, and manipulation to support operational analysis.

Skills: Database normalization, SQL (joins, views, stored procedures), ER modeling