Spatial analysis of crime often involves aggregating georeferenced crime data into areal units, generating a map of crime counts that can then be analyzed. Which areal unit should be used, though? This is an important decision, since changing the unit of area can impact analysis and results: the famous Modifiable Areal Unit Problem (Openshaw S, Taylor PJ, A million or so correlation coefficients: three experiments on the modifiable areal unit problem. In: Wrigley N (ed) Statistical applications in the spatial sciences, 127–144. Pion, London, 1979). In some cases, prior theoretical knowledge or even the research questions will recommend a specific areal unit (Weisburd D, Bruinsma GJ, Bernasco W, Units of analysis in geographic criminology: historical development, critical issues, and open questions. In: Putting crime in its place. Springer, New York, pp 3–31, 2009; Weisburd D, Groff ER, Yang SM, The criminology of place: street segments and our understanding of the crime problem. Oxford University Press, Oxford, 2012), but that might not always be the case. In this chapter, we present an approach that considers spatial properties of the original crime dataset to determine an adequate areal unit (see also Ramos RG, Silva BF, Clarke KC, Prates M, J Quant Criminol 37:419–443, 2021; Ramos 2019). Two properties are key: internal uniformity and robustness of the crime counts. If areas are too large, areal aggregation may hide crime clusters (low internal uniformity); if areas are too small, crime counts will be too prone to random fluctuations (low robustness of crime counts). For a more reliable crime map, these two factors need to be considered—the method presented in this chapter finds an areal unit that balances these two criteria. To illustrate the method, case studies using real-world crime data are shown at the end. A sensitivity analysis is also shown, illustrating how differences in some implementation details of the method affect the optimal unit but not to a radical extent, highlighting that the methodology identifies a range of acceptable granularities and the extremes that should be avoided.