Ezzatollah Mohammadi
H Şebnem Düzgün - Middle East Technical University (METU), Ankara, Turkey

Fire events in Ankara, Turkey, in years 2005 and 2006 from a statistical perspective have been analyzed, modeled and tested against Complete Spatial Randomness (CSR) to see whether data distribution exhibit random, clustering or regularity. To visualize fire event in Ankara, we first show the data in the form of dot maps and then for analyzing will go to open source software “R” and related packages, i.e. spatstat to trace the intensity and density of fire events in order to do the CSR tests. In a general point of view density of 200 in quadrat method are discussed and compared with other software to ensure of density behavior, then Ripley’s K-Function test are done to see the actual behavior of fire events. Concluding, from CSR tests we perceive clustering in fire events in residential and commercial area. It helps the decision makers to focus on such places for any action to control or reduce the fire events in those areas. We end the research with recommendations for future works.