A collection of data science and machine learning projects demonstrating skills in predictive modeling, optimization, deep learning, AI, and financial analysis. Projects include customer churn prediction, logistics optimization, bankruptcy prediction, emotion detection, and financial decision-making analysis.
Explored bankruptcy prediction in Italy's construction sector SMEs, focusing on imbalanced data. Python was used to implement and compare four analytical models, refining a methodology for broader application in economic studies, enhancing the precision of financial distress predictions.
Employed R to construct predictive models including Decision Tree, Random Forest, and Boosted Decision Tree to identify customers at higher risk of churn for a card-association company. The analysis was geared towards predicting the profiles of likely churners, enabling the deployment of tailored marketing strategies to effectively reduce churn.
Python was used to develop a deep learning model using Convolutional Neural Networks (CNNs) and classify facial expressions into seven categories. The project involved training and testing the model from scratch on the FER-2013 dataset.
Addressed a logistics challenge faced by JD.ID, focusing on minimizing transportation costs while ensuring demand fulfillment and adhering to capacity constraints. Utilized Pyomo, in Python, to evaluate and propose solutions across three different scenarios, effectively balancing cost-efficiency with operational demands.
Conducted a Python-based analysis to investigate the effects of political unrest on crucial economic indicators. Utilized four datasets to examine the relationship between political instability, oil price changes, and their impact on CPI and GDP. The study included a comparative assessment of two country groups, a control and a study group, to discern the economic impacts of political turmoil.