Last December, our GEO-C researcher, Mohammad Mehdi Moradi gave a talk entitled “introduction-spatial-point presentation” for master students in Geoinformatics at the institute for geoinformatik in the University of Munster, Germany. During his talk, he has discussed spatial point processes using different data examples. The lecture aims at providing students with simple definitions of point processes together with some statistical methodologies used to understand the behavior of data. Along with the definitions, applications have been shown by proving the R code as well. Thus, all examples in this lecture are reproducible if one is interested.

Generally, a ‘spatial point pattern’ is a dataset giving the observed spatial locations of things or events such as trees in a forest, road accidents, earthquake epicentres, etc. Each event might carry some extra information or labels (marks). For instance, the importance of analysing point patterns can be seen when studying crime data or traffic accidents. Using statistical methodology for point processes, high-risk regions/streets can be highlighted together with some significant reasons. Moreover, the relationship between different types of crimes/accidents can be revealed. These then might help Police department or Townhall in terms of city management and safety. For more details, see the slides and references therein.

This lecture covers the following topics:

1. Spatially varying distribution of the points. Is there any hot-spot? How events use space?
2. The type of interaction between points using second-order summary statistics.
3. How is the spatially variation of marks?
4. Relative Risk.

In this lecture, apart from simple definitions, each topic is attached with the corresponding R code using R packages spatstat and sparr.