A significant fraction of college students suffer from serious mental health issues including depression, anxiety, self-harm and suicidal thought. The prevalence and severity of these issues among college students also appear to increase over time. However, most of these issues often remain undiagnosed, and as a result, untreated. One of the main reasons of this gap between illness and treatment results from the lack of reliable data over time. While health care services in college campuses have been focusing on detection of illness onset and appropriate interventions, their tools are mostly manual surveys which often fail to capture the granular details of contexts and behaviors which might provide important cues about illness onset. To overcome the limitations of these manual tools, we deployed a smartphone based tool or unobtrusive and continuous data collection from 22 students during an academic semester. In this paper, we present the preliminary findings from our study about assessing mental health on college campuses using passively sensed smartphone data.