Data, Contextualisation, and Simulation Design: Findings from Pilot Interviews
This is the second research note by Onkar Hoysala, one of the researchers who received the Social Media Research grant for 2016.
In the second blog post for this series, I will briefly describe my motivation for carrying out this research, as well as some pertinent findings from the pilot interviews I carried out.
After completing my undergraduate degree in Computer Science and Engineering, I joined a research organisation which looked at infrastructure and policy studies, using gaming and simulation methods. A year and a half later, I was one of the co-founders of Fields of View, a research group using gaming, modelling and simulation methods to study cities. Over the course of the last few years, working on projects where computer simulations have played a crucial part got me interested in understanding where these technologies come from, who uses them, and how.
Being closely related, there has been a constant debate about the definitions of models, simulations, and games (Crookall, 2010) with some such as Crookall (2010) arguing that the debate is fruitless. However, for purposes of this work, it is useful at this point to lay down a working definition for a model, simulation, and game. I employ the following definitions:
- – A model is a representation of any real world phenomenon.
- – A simulation is the functioning of the model over time and space.
- – A simulation could be a computer-based simulation, with no human involvement; or could be a gaming-simulation, where human actors participate and play a role in determining the outcomes of the simulation.
In this project, I focus primarily on computer-based simulations. Epistemologically speaking, simulations are often thought to bridge the gap between theorising and experimentation (Winsberg, 2003). They form an apparatus through which to see the world from outside. However, a risk with simulations is that the users of this apparatus may not fully grasp its inner workings, thus giving the impression that simulations “offered a direct window onto nature” (Turkle & Clancey, 2009), or a one-stop quick solution to all their problems.
Having always been interested in transport simulations, and having been part of a group at Fields of View which worked on a multi-year transportation simulation proposal in collaboration with two universities, I was particularly interested in focussing on transportation simulations for this study. Further, exposure to different literature on anthropological works on simulation (for example, Turkle & Clancey, 2009), works on social construction of technology (see Bijker, 2009), and works on practice based studies of technology (for example, Orlikowski, 2000) got me interested in studying how technologies such as simulations are situated contextually. With this as the backdrop, I began working on this project, studying the practice of simulation designers in contexts such as India.
Before undertaking detailed field work through participant observations, it was important to first get a broad picture, or a ‘lay of the land’. To understand the same, I conducted 5 preliminary interviews with designers of simulations in transportation studies, as part of my pilot study. I interviewed the following people (names have been changed here for anonymity):
- – Vinay, a professor of civil engineering from a reputed university in Bangalore, whose group works on transportation simulation and policies;
- – Anirudh, a professor of civil engineering from a reputed university in Chennai, who works on traffic micro-simulation;
- – Jayanthi, a professor of civil engineering from a reputed university in Chennai, who works on traffic models and traffic data collection, and whose projects have been demonstrated for the traffic police;
- – Chetan, the founder of a policy think tank in Tumkur district, who has worked in, and with, government organisations in Bangalore;
- – Bhargav, one of the co-founders of a research organisation in Bangalore, who develops computer based simulations in transportation, among other areas.
To analyse these five interviews, I followed a two-cycle open coding method (Saldaña, 2012) to derive codes and categories for analysis. In the First Cycle of the coding, I coded the five transcripts with words and phrases which best described the responses. In the Second Cycle, I reconfigured the codes to remove codes with the same meaning, created a coding tree, and created categories for analysis.
The diagram below represents the categories that evolved from the coding.
Figure 1 Categories of Analysis of the preliminary interviews
The most pertinent findings from the preliminary interviews were about the kind of simulation work that the groups here do, and the most pressing issues they have in carrying out the work.
Contextualisation and Translation
The software used for designing transportation/traffic simulations are typically designed and developed in Europe (for example Vissim) or the USA (for example Transims). From the interviews, it is evident that a lot of the work by transportation simulation designers in India involves contextualisation of these tools, to calibrate and validate them for Indian contexts. Contextualisation involves modifying model parameters for lane discipline, for example.
The work on contextualisation has significant impact academically, through the development of new transportation sysetms models, or traffic flow models for Indian traffic conditions, or road design guidelines. For example, Vinay’s group has
“defined some new queueing methods . . . we called it PFIPFO – Probably First In Probably First Out, as the queue discipline we see at our intersections. We have incorporated this aspect into coming up with new delay models to estimate delay of vehicles at intersections, to optimise the traffic signal timings in a better and more effective and efficient way.”
However, translation of the theoretical work into practice – into a real world demonstration or into policy – has not often been possible. Chetan opines that while there is significant analytical depth in the simulation models, the tools are “very nascent” in their development and use in Indian social-political scenarios. Translation is so rare that when it does happen, it becomes something
“we can kind of boast about, [as] we may be the only academic institution to go into the field, implement it in the field and make it happen”.
Issues of translation arise due to other multiple, interrelated, issues such as inter-agency collaboration, or importance given to simulations by policy makers. One issue, however, stands out, with each of the respondents discussing it as critical – Data.
There are multiple aspects to the issues regarding transportation data, and I here I only touch upon three key aspects: Data access, Data collection, and Data integrity.
“The modelling was not very strong here [in India], mainly because there was not enough data available. There was no automated system that could give you continuous data . . . students would go and enter data manually, come back and fit models.”
Jayanthi’s statements are echoed by other practitioners as well, with respect to availability of data. In order to overcome lack of data, Vinay’s group designed and carried out a primary survey to collect travel activity data using travel diaries, and published an article describing the contextualised travel diary.
Even if data is being collected, having access to them is yet another issue, with government data collection agencies unwilling to share them. For example, while there is data being collected about various aspects of transportation, such as GPS locations of certain buses, Chetan says:
“there are issues of sharing and access, and in my experience, none of this data is being put to use to develop realistic models for simulation, or being shared easily”.
Further, with multiple agencies such as Public Works Department, Traffic Police, Ministry of Transport etc., being involved, each of them collect data individually, with “none of them being reliable”, as the data collection methods are manual. The problems are not just lack of automated systems, but a lack of systematic collection of data itself, or the problem having the necessary technologies but not effectively utilising them. A common example from the interviews is that of video camera data from traffic intersections. The traffic management system in Bangalore has “an excellent set of technologies . . . They should be lauded for doing that. But it is all minus intelligence”, according to Vinay.
Apart from data that is collected by the State, transportation simulation designers do gain access to data from private organisations (such as mobile phone companies) through their engagement with such organisations as part of project consortiums. Vinay, for example, is collaborating with one of India’s largest mobile phone network operators to get cell phone data, using which traffic could be mapped in any area where mobile phone coverage is available.
While the respondents primarily focussed on contextualisation and issues related to data, aspects such as collaborations with different stakeholders; institutional frameworks related to transportation; and what they think “impact” means – were other areas of discussion, which deserve further enquiry.
Going forward, I will be following up with each different members of each of the groups with detailed interviews. While these interviews may provide a detailed picture of the context, a practice-based study requires an ethnographic method. Due to limitations of time, a thorough ethnography of simulation software use may not be possible, but I plan to spend around 4 weeks (a few days a week) with one identified group to study their use of simulation software to design contextualised simulations for Indian contexts.
Bijker, W. E. (2009). Social construction of technology. A Companion to the Philosophy of Technology, 88-94.
Crookall, D. (2010). Serious games, debriefing, and simulation/gaming as a discipline. Simulation & gaming, 41(6), 898-920.
Orlikowski, W. J. (2000). Using technology and constituting structures: a practice lens for studying technology in organizations. Organization science, 11(4), 404–428.
Saldaña, J. (2012). The coding manual for qualitative researchers. Sage.
Turkle, S., & Clancey, W. J. (2009). Simulation and its Discontents. Cambridge, MA: MIT Press.
Winsberg, E. (2003). Simulated experiments: Methodology for a virtual world. Philosophy of science, 70(1), 105-125.