International Journal of Remote Sensing Applications Volume 4 Issue 2, June 2014 http://www.ijrsa.org doi: 10.14355/ijrsa.2014.0402.03 97 Pedestrian Safety Modelling and Analysis Using GIS in Chennai Rohith P Poyil**, Anil Kumar Misra**, R Murugasan*** **Department of Civil and Environmental Engineering, ITM University, HUDA Sector 23 A, Gurgaon 122017, India ***Institute of Remote sensing, Department of Civil Engineering, Anna University, Chennai 600025, India rohithpoyil@itmindia.edu; anilmishra@itmindia.edu; murugasanr@rediffmail.com Received 29 th March 2014; Revised 31 th March 2014; Accepted 16 th April 2014; Published 03 rd June 2014 © 2014 Science and Engineering Publishing Company Abstract This paper analyzes the various variables affecting pedestrian road crashes, placing emphasis on the effect of daily activity patterns and the built environment, including the land use of the places. Study also develops a level of safety model in terms of number of pedestrian accidents in Chennai. First, eleven potential factors influencing pedestrian level of safety are summarized: width of road, width of sidewalk, average running speed of vehicles, vehicular volume, pedestrian volume, percentage of sidewalk area encroached, presence of crossing facilities, sight distance, pedestrian refuge and median, lightings and curb. The selected roads are typical of those prevalent in urban areas of Chennai. With the survey data, a stepwise regression analyses are carried out to develop a reliable pedestrian level of safety model for road segments, suitable for use in the vast majority of Indian urban areas. The study reveals that the factors significantly influencing pedestrian level of safety at road segments including width of sidewalk, average running speed of vehicles, percentage of sidewalk area encroached, presence of pedestrian refuge and median, lightings and curb. A model to predict the pedestrian safety level in terms of number of pedestrian accidents is developed in this study using regression analysis and integrated with GIS to produce a colour coded maps showing the predicted number of pedestrian accidents. The validation has been done by comparing the predicted number of pedestrian accidents with actual number of pedestrian accidents occurred. Keywords Safety Model; Stepwise Regression Analysis; Pedestrian Level of Safety; GIS Introduction Worldwide pedestrian-vehicle crashes are considered as one of the major public health concern, but it is a preventable cause of death and hospitalization. Approximately 400,000 pedestrians lost their lives in pedestrian-vehicle crashes annually (Naci et al., 2009) worldwide. Studies carried out by several researches revealed that almost half of all pedestrian fatalities in Europe involved elderly pedestrians (Hakamies- Blomqvist, 2003; Organisation for Economic Co- operation and Development, 2001). The potential safety importance of the walking direction along a road can be done by examining pedestrian accidents as a function of exposure to risk (Juha et al. 2013). Research on user behaviour and preferences has been a helpful tool in improving road safety and accident prevention in recent years (Mario et al. 2014). A systemic inquiry and the transformation of the urban road network should be performed in order to diminish the areas of vehicle–pedestrian conflicts and to significantly reduce vehicle speeds in areas of pedestrian presence and activity (Victoria et al. 2012). Walking is a basic human activity and pedestrians are a part of every roadway environment. Everybody is a pedestrian at one point or another. A person’s decision to walk or use other modes of transportation is greatly influenced by his safety concerns. Pedestrian Level of Service model for arterials represents a progressive shift in evaluating the quality of service from a provider-based measure to a user-based measure (Theodore et al. 2004). With the full sample of collisions, binomial logit models estimated the odds of collision occurrence as related to the road and the neighbourhood environments and adjusting for exposure (Anne et al. 2008). Spatial analysis techniques can be combined with regression models to understand factors associated with risk (Robert et al. 2004). Geographic Information Systems (GIS) based approach can be used to compare the average walking