Information on stays and days spent in shelters is among the most basic of data collected through administrative records of homeless facilities. This paper goes over methods by which to analyze this data and which yield results that contribute to understanding the dynamics of how people use shelters. First, the paper shows ways to convert shelter stays, however they may be measured, into standardized units called episodes. The second section covers measures of central tendency and frequency distributions, methods that are likely to be readily familiar to most persons. The third section explains two basic methods of survival analysis, survival curves and hazard curves, while the fourth section describes regression procedures, which are more complex measures of survival analysis. Finally, the fifth section covers two methods for segmenting the shelter population by use pattern: the basic heavy user analysis and the more complex cluster analysis. In surveying all of these methods, frequent examples are provided to illustrate the concepts that are discussed and to facilitate understanding of how the specific method of analysis might be used to provide practical information that would be of use in operating and improving shelter services.