Interactions between people and natural spaces, through leisure or tourism activities, form a complex socio-ecological spatial network changing over time and space in response to economic, social, political and cultural drivers. In the age of information, the increasing availability of large databases generated by the use of geolocated information and communication technologies (ICT) devices allow us to gain a better understanding of complex socio-ecological interactions at an unprecedented spatio-temporal resolution. Within this context, we propose to model and analyze interactions between humans and natural areas based on information extracted from geotagged photographs embedded into a multiscale socio-ecological network. We apply this approach on 16 case study sites in Europe relying on a Flickr database containing hundreds of thousands of validated and classified geotagged photographs. After evaluating the representativeness of the network, we investigate the impact of visitors' origin on the distribution of socio-ecological interactions at different scales. At a global scale first, clustering the sites according to a spatial measure of attractiveness, we detect four groups of sites. Then, at a local scale, we investigate how the distance traveled by the users to reach a site affects the way these users interact with this site in space, time and landscape diversity. The approach developed here, integrating ICT data in a network-based framework, offers a new way of visualizing and modeling interactions between humans and natural areas for the development of efficient ecosystems conservation and protection planning strategies.