Analysing point patterns on networks — A review

We review recent research on statistical methods for analysing spatial patterns of points a network lines, such as road accident locations along network. Due to geometrical complexities, the analysis data is extremely challenging, and we describe several common methodological errors. The intrinsic lack homogeneity in militates against traditional statistics based stationary processes. Topics include kernel density estimation, relative risk parametric non-parametric modelling intensity, second-order using K-function pair correlation function, point process model construction. An important message that choice distance metric pivotal theoretical development real data. Challenges computation are discussed open-source software provided.