Existing epidemiological models analyze the transmission of infectious diseases considering a perfect homogenous population. However, the COVID-19 emergency has shown the importance of considering activity-travel behavior when studying the spreading of the virus. With increasing epidemiological data available, and the outburst on agent-based activity models, we can move beyond aggregation and start including individual features. To the best of the authors knowledge, this is the most in-depth study of how socio-economic and virological features impact the spreading of COVID-19 to date. We use chronological data from the Federal Office of Public Health (FOPH) from mid-February 2020 to midSeptember 2021. We derive the influence of socio-economic characteristics with a novel semidisaggregated SIRD model, obtaining the total number of infections per specific group of the population sharing pre-determined features. Finally, we validate our model with Google data and compute the reinfection rate by applying non-pharmaceutical interventions. Five features, including information about the individual and the municipality, have a ≥ 95% probability of being correlated with the endogenous variable of positive testing for COVID19. In addition, we find that certain variables, including age or the population density per square meter, remain representative for all waves, whereas others, like household income, are dependent on the epidemiological wave studied. Our results suggest a strong dependency on individual and municipality characteristics and the force of infection of COVID-19.