Invasive candidiasis is associated with high morbidity and mortality in critically ill patients, i.e. patients admitted to Intensive Care Units (ICUs) or in surgical wards. There are no clinical signs or specific symptoms and even though early diagnosis risk scores and rapid tests are available, none of such strategies has an equally-optimal level of sensitivity and specificity. In the era of Electronic Health Records (EHRs), several clinical studies exploited Machine Learning (ML) models and large database of features to improve the diagnosis accuracy. The main aim of this work is to build a wide dataset which can be exploited to apply ML models to further improve the early recognition of candidemia at the bedside of patients with compatible signs and symptoms.