Análise Exploratória e Preditiva dos Dados do Titanic: Identificando Fatores de Sobrevivência.

Abstract

This study aims to analyze the Titanic dataset using Exploratory and Predictive Analysis techniques, seeking to identify the most relevant factors for passenger survival. The methods applied initially involve Exploratory Data Analysis (EDA), including missing value inspection, univariate and bivariate analysis, correlation studies, and statistical visualizations to understand the patterns present. In the predictive stage, Logistic Regression, Decision Tree, and Random Forest algorithms were used, implemented in Python using the VS Code development environment. These models were applied after data preprocessing, including variable standardization, creation of dummy variables, and division into training and test sets. The results showed that the Random Forest and Logistic Regression models presented the best predictive performance, achieving higher accuracies than the other approaches. The analyses revealed that variables such as sex, ticket class (Pclass), and age are among the most influential factors in the probability of survival. Furthermore, Random Forest demonstrated a greater ability to identify non-linear relationships between variables. It is concluded that the combination of exploratory and predictive techniques allows for a broad and well-founded understanding of the dataset, highlighting the importance of EDA for efficient modeling. The study reinforces the value of machine learning methods in identifying patterns and supporting decision-making, demonstrating the potential of using Python and computational tools in academic and professional analytical practice.