Social Vulnerability Segmentation Methodology Based On Key Performance Indicators

Detecting social vulnerability is a complex process that involves many personal and environmental factors to take into account. In this paper, a new methodology is proposed to detect which individuals are more likely to live under vulnerable situations merging all those personal and environmental factors in one metric. This methodology assigns each person who asks The Red Cross for help different normalized key performance indicators (KPI) regarding several areas as personal, social, health, environmental, economic, labour and familiar. All these fields are combined to represent the person in two axes: the internal or personal vulnerability axis (composed by personal, health and economical indexes) and external axis (composed by environmental, labour, social and familiar indexes). Based on these axes and through unsupervised machine learning techniques, this methodology assigns each per-son a vulnerability group which may be related with a series of actions to cover their needs and act upon their situation. This way, the proposed methodology allows us to go from a high dimensionality to a reduced problem that considerably simplifies the study. This process permits The Red Cross act quicker in those high-vulnerable or more-likely-vulnerable situations, improving the assistance process and helping the estimation of needed re-sources.