Follow Up on Ethnographic Study of Taxi Drivers' Use of Location-Aware Systems

Following up on the preliminary work to setup an ethnographic study of the use of GPS navigation systems by taxi drivers in Barcelona. I drilled the taxi driver who brought me to BCN El Prat airport at 5am this morning. He was at the end of his night shift. I started by trying to define his profile. He is very familiar with the city and has been a taxi driver for 7 months. He bought his GPS navigation system (that he did not use for this ride) with the taxi. While he admitted that GPS (he was saying “GPS” instead of “navigation system” does not come without problems (bad advises on navigation) and he considers it as a killer asset. When I encouraged him in telling me more about the necessity to be aware of his location and his whereabout, he detailed two different scenarios:

1. Reaching the specific destination
During the ride to a rather unknown destination (lack of precise mental model), he accesses the geospatial information as in a “funnel”. First, he checks in the paper-based guide of the city to know more or less the area he should direct to. In that area, he waits for a traffic light to type the exact address in the navigation system. He explained that he engages with the system at that moment to avoid the often misleading information on the path to take. Indeed, a taxi driver applies several paths depending on the time of the day and circumstances (e.g. traffic, weather conditions, recommendation of passenger). In fact, this specific taxi driver could name me the places where he experienced the navigation information to be absolutely irrelevant (e.g. get access to the forum area, plaza de Glories).

2. Getting back on tracks
A main problem for a taxi driver is to be lost after dropping off a customer. This often happens when the customers guide to the destination. Therefore, taxi drivers thrive on information to get back into known terrotory. That can be landmarks (e.g. a tall building), the topology (e.g. a mountain/hill). But getting a sense of orientation can easily become problematic (e.g. during nighttime or bad weather conditions when tall buildings and montains are not clearly visible). This is where the GPS comes handy. This specific taxi driver would ask the them to give him the way out by typing “plaza espana” (i.e. a large roundabout in the center of Barcelona).

From that informal, unstructured discussion, I tried to form a more detailed description on what I am looking for in the study (questions/hypothesis) and how to answer (methodology/data analyis)

What are the hypothesis/questions

1. When is the system used? That is at when does the driver need it during a service (start, to complete the task, at a special event (find related information such as POI or hotel).
2. Where is the syste used?: dense urban area, countryside, main roads, familiar zones, unknown remote areas.What kind of geoinformations are used by the drivers both from the system and from the environment (experience, radio, interaction with customer, context such as visibility to landmarks). First I am interested by the nature of the information (step-by-step navigation, real-time positioning with map to get a sense of the whereabouts, POI, dynamic data such as radars/traffic, time of the day. This actually reminds me of a taxi drivers who explained me that he essentially used GPS to relax. That is, he would forget about the names of the streets and the intersection (or count the number of the streets) because he could always fallback on his real-time position and act based on that. Second, I am interested to record how the drivers rely on these data/observations considering that some are uncertain.
3. What type of view is used and when (top, titled (bird’s eye),.voice)? What is the preferred view (from part of the taxonomy for visualizing location-based information) depending on the specific tasks. That is to understand what views help disambiguate or support the awareness of the whereabouts.
4. How does the co-evolution process unfolds? Maybe drivers can explain the evolution of their relation with the system. From the acquisition and setup to the point of clearly defining what the system can and cannot deliver. I am especially interested in knowing how the drivers learned the limitations of the system. How much the system can be trusted and what is the reaction when the quality is not met (awareness of the limitations/imperfections)

What is the methodology to collect data/record data
I plan 40-50 semi-structured interviews of taxi drivers at their waiting area in the BCN airport. I would then transform myself into a complete observer (focused ethnography) in about 10 rides of 30 minutes. field notes and photos of interaction. To be detailed more as I get more familiar with qualitative data collection. I should also be able to argument for that approach and not another. The BCN airport has a huge parking lot where taxis wait their turn. The minimum time is 1h. Taxi drivers take the opportunity to clean, walk, eat and play. I was confirmed it is a very good place to conduct survey with semi-structured questionnaires.

Taxis queuing area at Barcelona airport (often for more than 1 hour). The stage for semi-structured interviews. (photo courtesy of jpg)

How to conduct the data analysis
Coding: Categorize the unitary tasks (navigation (take decisions on the path), search destination, self-location awareness (positioning) of the taxi drivers and understand what type of geospatial information is applied for each. How are the limitations/uncertainty understood (if there are any), what was the evolution process as of defining the limits of the systems.

Next step: Maybe focus on 2-3 questions and define a set of leading questions to structure an interview. Try to grab Ken Anderson and/or Barry Brown at Pervasive to get some insights. I hope to get some more insights from the Ethnography, Thick and Thin tutorial on Wednesday.