Projects
M.Sc. Thesis – Causal Machine Learning for Survival and Quantile Treatment Effects
Keywords: Causal Inference, Survival Analysis, Right-Censoring, Efficient Influence Functions, Machine Learning
Master’s thesis developed within the M.Sc. in Stochastics and Data Science at the University of Torino.
The work proposes efficient, data-adaptive estimators for treatment-specific survival curves and quantile treatment effects under right-censored data, combining causal inference theory with modern machine learning methods.
Supervisors:
- Matteo Giordano (Supervisor)
- Stijn Vansteelandt (Co-supervisor)
Main contributions:
- Identification of causal survival and quantile treatment effects under dependent censoring
- Cross-fitted one-step estimators based on efficient influence functions
- Data-adaptive nuisance estimation via Super Learner ensembles
- Theoretical guarantees (consistency, asymptotic linearity, inference)
Thesis: Efficient Estimation of Survival Curves and Quantile Treatment Effects under Right-Censored Data: a Causal Machine Learning Approach
Revenue Forecasting at Italdesign
Keywords: Time Series Forecasting, LSTM, XGBoost, Applied Machine Learning
Designed an end-to-end revenue forecasting pipeline combining deep learning and gradient boosting methods.
The project focused on feature engineering, model comparison, and robust evaluation for decision support in an industrial context.
Main contributions:
- Hybrid forecasting models based on LSTM and XGBoost
- Systematic evaluation of forecasting performance metrics
- Deployment-oriented pipeline design for business analytics
Code: Private repository
Portfolio Clustering at Italdesign
Keywords: Unsupervised Learning, Clustering, Business Analytics, scikit-learn
Applied unsupervised learning techniques to cluster and analyze the company’s internal portfolio.
The objective was to identify meaningful segments within historical data to support strategic planning and revenue analysis.
Main contributions:
- Data preprocessing and feature extraction for portfolio analysis
- Clustering and dimensionality reduction techniques
- Visualization of portfolio segments and strategic insights
Code: Private repository
Data Analysis on COVID-19
Keywords: Data Analysis, Epidemiology, Time Series, Python, Pandas
Conducted an exploratory and statistical analysis of COVID-19 epidemiological data to study temporal trends and regional patterns during the pandemic.
The project focuses on data cleaning, visualization, and quantitative analysis of publicly available datasets.
Main contributions:
- Data preprocessing and integration of heterogeneous COVID-19 datasets
- Time series analysis of cases, deaths, and testing dynamics
- Visual analytics to highlight temporal and geographic trends
Code: github.com/Edoishere/DataAnalysisOnCovid19
Additional Projects
Further projects, experiments, and code samples are available on my GitHub profile:
https://github.com/Edoishere