Caracterización de las infracciones de tránsito en Bogotá en el año 2017 utilizando un modelo Cox Log - Gaussiano de patrones puntuales

Abstract

The spatial analysis of point patterns is used to know the behavior of a set of data throughout a spatial observation window, and then to propose models that allow making predictions about the number of occurrences of a phenomenon in the place of study. In this work, an extension of these analyzes is made, where not only the spatial behavior is studied, but its dependence through time. For this, a database of traffic violations that occurred in the city of Bogota during the year 2017 is used, but the violation with the greatest occurrence during that year was selected, in this case, the traffic violation C02, which corresponds to parking in prohibited places. For the analysis, several options are considered: first, observe purely spatial behavior without including time as an analysis variable, working time as a qualitative mark, a quantitative mark and finally, as one more dimension. It was concluded that the latter is the most optimal option for the analysis, finding that both at a spatial level and at a space-time level, there is an added pattern . Finally, we have the relative risks of the model, making forecasts at some times chosen at random.