Electricity is essential for the functioning of modern society, underpinning both economic and social sectors. A reliable power system depends on efficient generation, transmission, and especially distribution networks, since failures in these systems can lead to serious socioeconomic consequences. The primary distribution network, operating at medium voltage, connects substations to consumers and plays a central role in ensuring continuity and quality of energy supply, which is monitored through indicators such as SAIDI and SAIFI. The vulnerability of these networks may result from aging infrastructure, increasing demand, adverse weather conditions, operational failures, or interactions with vegetation. This study aims to identify the main factors contributing to the fragility of primary distribution networks through data analysis. The research encompasses the collection and organization of historical failure data, the application of statistical analysis to identify principal agents of vulnerability, and the development of a statistical classification model utilizing logistic regression. The results highlight significant contributing factors, develop a predictive failure model, and offer both seasonal and spatial analyses, thereby facilitating the visualization of regions most vulnerable to disruptions. This approach supports more informed decision-making in grid management and maintenance, helping to prioritize investments and mitigate failures. Ultimately, this work contributes to the robustness, resilience, and continuity of electricity supply, ensuring better reliability and quality for end users.