Criminal hotspots are heuristically understood, but seldom well-defined and empirically evaluated. In this thesis, I examine the concentration of crime into microgeographic hotspots, testing both the extent to which this occurs across major cities and the relationship between spatial features and crime. I find that roughly five percent of street segments are responsible for half of crime across major cities, with this concentration level being robust to changes in total crime rate and economic conditions over time. I also find a significant relationship between the presence of spatial features such as nearby schools, bus stops, bars, and graffiti with the crime level in microgeographic units. Through a routine activity and crime pattern theoretic interpretation, such spatial models of crime can help to identify features and facilities that attract, inspire, and deter crime. These findings have policy relevant implications for both urban planning and police strategy, offering intuition as to where crime can be expected to concentrate and how changes to local environments impact public safety.