Circular Statistics for Directional and Temporal Data : Case Studies of Lineaments and Noise Violations Offence

Abstract

Circular statistics or directional analysis are required for appropriate processing and analysing of circular spatio-temporal data, including various angular measurements such as trajectory data, and temporal cyclic data such as time of day or day of week. Suitable SW tools offer several types of plots and diagrams and enable calculation of measurements of central tendency, variability and concentration, and to test distribution and differences between circular variables. A short review of basic statistical indicators and usual tests is provided. The application is demonstrated on two case studies. Directional analysis of morphological lineaments and old faults show partial coincidence of directions, but also document how basic circular statistical indicators are influenced by bimodal or dispersed distribution. Multisample testing confirms significant differences between orientation of faults and lineaments. Analysis of noise violation offences in Ostrava provides an overview on how to process cyclic temporal data. The time of day is highly concentrated and depends on the day of week. To the contrary, the month and day of the year show uniform distribution.