International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Pedestrian route choice modelling at a large scale mega religious gathering
Tarapada Mandal, K Ramachandra Rao, Geetam Tiwari

Last modified: 28 March 2017


Pedestrian motion is a complex process which happens as the outcome of many behavioural decisions. These decisions can be defined at different levels. Route choice is a key behaviour in any kind of activity making. The route choice at tactical level can have severe consequences at the operational level where the pedestrians actually perform the activity. So route choice of pedestrians in any activity is a prominent behavior that should be understood well to avoid any kind of mishaps. Knowledge of human decision process in a complex environment helps crowd managers in handling any situation. If not managed properly small incidents may lead to crowd disasters like stampede. Several past events of stampede were found in the literature including religious events in India (Illiyas, Mani, Pradeepkumar, & Mohan, 2013). This study focuses on understanding the route choice behavior of the pedestrians who partook in the world’s largest religious gathering (Kumbh Mela) with the number of pilgrims estimated over 50 million in 2016. The large crowds throng the mela ground to take a dip in the holy water of the river Kshipra in Ujjain city which is situated in the state of Madhya Pradesh in India. They also visit several akharas (temporary camps of Saints) to listen to religious preaches delivered by the saints. In this course of religious activities, many temporary and permanent routes are used by the pilgrims. As Kumbh Mela encompasses a huge mass gathering, it is very important to understand how these large gatherings are influenced by several attributes of the routes.


The objective of this study is to find the key attributes that influence the route choice behavior of pedestrians. Further, the level of impact of these attributes on the route choice behavior of the pilgrims is to be studied. The study objective also includes the understanding of the diversities in the choices of pilgrim groups with distinct socio-demographic characteristics in the complex environment of mass gathering.


In addition to some studies on the pedestrian movements at large gatherings (Galama, 2015; Guo & Loo, 2013), discussion with the experts and the well-travelled pilgrims, this research takes into account ‘crowdedness’, ‘shortcut’, ‘width of the routes’ and ‘proximity to water of the routes’ as the main attributes in route choice of pilgrims during the mass event. Further, literature also discusses the use of discrete choice models to imitate route choice behavior of pedestrians in different settings. Based on the review of literature it is evident that the route choice behavior of pedestrians’ at large religious gatherings has not been studied in detail. The salient features of this study are:

i)                    Data Collection for pedestrian route choice using a stated preference survey at a religious mass event (Kumbh Mela) using in vivo interviews.

ii)                  A partial factorial design is used to reduce the number of questions to avoid any kind of agitation of the pilgrims. Binary method is used to form the factorial design for the questionnaire. Socio-demographic information about the interviewees is also recorded.

iii)                Use image based technique (Harline & Burris, 2014) to capture the choice making of pedestrians enabling them to visualize well the context of the environment before making any choice.

iv)                Use a binary logit model (as binary method is used to quantify the attribute levels in the factorial design) to mimic the decision making process of the pilgrims.

The results of the Pearson correlation test show that there is a correlation between the socio-demographic characteristics of the pilgrims and their choices. Several models (binary logit) are run to estimate the influencing parameters of pedestrian route choices. Biogeme 1.8 (Bierlaire, 2003), a software is used in this study. The estimated final log-likelihood and the adjusted rho-square values obtained in this study suggest a good fit for the data used in the study.The parameter estimates of the different attributes indicate that pilgrims are less likely to go to the crowded areas. They are more likely to take shortcuts and prefer to avoid routes through narrow alleys. The results also suggest that proximity to water has a positive impact on the route choices of pilgrims i.e., they are more willing to choose routes near to the water (or some routes through the temporary pool made on the river). However, a better estimate of the attributes of the routes could be attained by using virtual reality method instead of image-based stated preference method but the reliability related issues of  virtual reality technique (as it has low ability to diminish the generalization of all visual parameters) in this context is to be resolved before using the method.



Bierlaire, M. (2003). BIOGEME: a free package for the estimation of discrete choice models. Swiss Transport Research Conference. No. TRANSP-OR-CONF-2006-048.

Galama, I. M. (2015). Route Choice Behaviour at Mass Events. Masters of ScienceThesis. Department of Transport and Planning. Delft University of Technology. Retrieved from

Guo, Z., & Loo, B. P. Y. (2013). Pedestrian environment and route choice: evidence from New York City and Hong Kong. Journal of Transport Geography, 28, 124–136.

Harline, C., & Burris, M. (2014). Impact of Traffic Images on Route Choice and the Value of Time Estimates in Stated Preference Surveys. Transportation Research Record: Journal of the Transportation Research Board, 2405, 24–32. article.

Illiyas, F. T., Mani, S. K., Pradeepkumar, A. P., & Mohan, K. (2013). Human stampedes during religious festivals: A comparative review of mass gathering emergencies in India. International Journal of Disaster Risk Reduction, 5,10-18.

Conference registration is required in order to view papers.