A longitudinal analysis of the COVID-19 effects on the variability in human activity spaces in Quito, Ecuador

The COVID-19 pandemic has had a huge impact on human activities due to lockdowns or travel restrictions to preserve public health and decrease the workload of hospitals. Therefore, human activities spaces (HASs) were deeply affected worldwide, but to an extent that is hard to quantify properly. This paper presents a longitudinal analysis of HASs in Quito, Ecuador, before and during the COVID-19 pandemic. Using location data collected through Google Location History (GLH) from the Google Maps application, we compute weekly people's activity point locations (APLs) from a convenience sample of 263 participants, mainly composed of university staff members, considering only weeks with at least five days of data. These APLs are then used to measure the HASs using the confidence ellipses and the minimum spanning trees. Finally, we perform a weekly intra-personal and inter-personal variability analysis of the HASs using a random intercept model, considering (a) the size of HASs as the dependent variable and (b) the levels of restrictions due to the pandemic and the participants' demographics as independent variables. The results reveal that HASs are strongly affected by the intensity of non-pharmaceutical interventions (NPIs) (Social distancing, quarantines, lockdowns, travel restrictions or closure of schools and workplaces) and the composition of the socio-demographic groups. We also demonstrate that the disruptive effects of NPIs on human mobility were reflected in the decrease in trip durations in conjunction with a drop in visited locations as individuals only engage in essential neighbouring activities, implying substantial variations in the size and extent of HASs.