Federico Giorgi

Projects

• Optimizing Seattle Bus Transport Routes - 2025

  Urban congestion presents economic, environmental, and social challenges. This project leverages graph theory to optimize Seattle's public transit system by designing an efficient bus network. The approach involves modeling the city's road infrastructure as a weighted graph, constructing a Minimum Spanning Tree (MST) to minimize redundancy, and refining routes through optimization techniques. The model ensures better transit coverage, strategic bus stop placement, and improved accessibility while maintaining operational efficiency.

• Asian Portfolio Optimization - 2025

  This project analyzes ETFs representing the “Four Asian Tigers” (Singapore, South Korea, Taiwan, Hong Kong) using financial statistics and portfolio theory. It examines risk-return tradeoffs, correlation, VaR, and constructs efficient portfolios with and without short-selling. EWT (Taiwan) emerges as the best-performing asset, though diversification benefits are limited due to high correlation.

• IBM Employee Attrition - 2024

  Using HR data on over 1,400 employees, this project applies preprocessing, feature engineering, and multiple ML models (Logistic Regression, Random Forest, MLP) to predict employee attrition. The analysis includes model tuning, feature importance, and policy implications, focusing on actionable strategies to reduce turnover.

• Credit Scoring - 2024

  This Machine Learning group project aims to develop a predictive model related to financial difficulties among borrowers using various Machine Learning techniques. The project seeks to mitigate the issue of imbalanced data and is based on multiple modeling approaches, including Logistic Regression, SVMs, Gaussian Naive Bayes, KNeighbors Classifier, and ensemble methods.

• Solow-Swan Growth Model R Analysis - 2024

  Econometric analysis of GDP per capita using World Bank data. The project tests and expands the Solow-Swan model by incorporating R&D expenditure and institutional quality (Rule of Law), improving explanatory power and meeting OLS assumptions through data filtering.

• Simulated Annealing - 2024

  This project involves discrete landscape optimization using simulated annealing versus a greedy strategy. It explores the landscape defined by a discrete function represented by a matrix, where each entry holds the value of the target function for two inputs (column and row). The objective is to minimize this function and compare the performances of both strategies.

• Analysis of AirBnB 2022 dataset - 2023

  Analysis of the AirBnB 2022 dataset to identify the variables that affect the price in the six biggest USA cities.