After years of research on the analysis of individuals by automatic methods, the computer vision community has transferred its attention on the new issue of modeling gatherings of people, commonly referred as groups or crowds, depending on the number of people involved. The aim of GROW 2015 is to bring together a wide range of researchers in computer vision and machine learning from one side, and applied social sciences on the other, to share innovative ideas and solutions for exploiting the potential synergies emerging from the integration of the two domains, for a range of different applications. For this reason, the invited speakers of the workshop will come from the computer science and from the social science domains, promoting an intriguing cross-fertilization among the two areas. This will serve to get an answer to many unresolved issues in the computer vision community, like: what is a group and what is a crowd? Is it true that the difference among them is a matter of number of people involved? Are there cues other than the spatial proximity and oriented velocity which could be used to detect them? Are there different typologies of groups and crowds? For example, in the sociological literature, social studies defined different kinds of crowd (spectator crowd, casual crowd, protest crowds and some other) and it could be interesting to understand whether these definitions could have some computational counterpart. Are there issues for sociologists which could be faced using computational methods (like annotations, massive observations by ecological camera networks)? The aim of the workshop will be that of explicitly managing these issues, having also a panel discussion at the end of the event.