Potential dissemination of epidemics based on Brazilian mobile geolocation data. Part I: Population dynamics and future spreading of infection in the states of Sao Paulo and Rio de Janeiro during the pandemic of COVID-19.

Pedro S. Peixoto; Diego R. Marcondes; Cláudia M Peixoto; Lucas Queiroz; Rafael Gouveia; Afonso Delgado; Sérgio M Oliva

In order to asses the mobility patterns in weeks following the first cases of COVID-19 in Brazil in March 2020, we always take the mobility in March 2019 as a control group. Indeed, we need a measure of the usual mobility between the regions to compare with the observed mobility to know if it is within the usual pattern. For this purpose, we disregard the first days of March 2019, as the mobility was influenced by a major Brazilian holiday, the carnival week, so we observe the pattern of mobility in March 2019 starting on the 11th. On the one hand, the mobility in March 2020 is measured daily, by the proportion of movement from one region to another, i.e., by the daily transition matrices. On the other hand, the mobility in March 2019 is measured by the mean of these proportions over all considered days of March which fell on a day of the week, i.e., for each day of the week we calculate the mean of the proportions for all considered days of March 2019 which fell on it. Proceeding in this way, we have one transition matrix for each day of March 2020, and seven transition matrices related to the mean pattern of movement of each day of the week in March 2019. Each day of March 2020 is compared with the pattern of the day of the week it fell on. The analysis of this study concentrates on an important feature of the pandemic spread, that is, possible focus of future infections. We now discuss how they can be evaluated from the available mobile data.