In the first part of this case study, I explained the principles of the schematic map used to represent Paris’ transit system. Schematic diagrams are a popular and efficient way to map the transit system of urban cities. However, with the increase in the number of stations and the rapid development of infrastructure, it got me wondering if this representation is still up-to-date and if it is the best way to help passengers navigate…
In theory, public transportation should be really simple. Read a map, plan your route and occasionally maybe change trains. The problem is in practice it just doesn’t feel that simple.
Commuting in Paris can be hectic as hell sometimes and I speak from experience. This is typical for major urban cities around the world and I get my fair share of complaints from my colleagues in the Berlin and Toyko studios. However, Paris is particularly tough when it comes to public transport and communicating, as it is a relatively small and dense city. This often leads to traffic congestion, resulting in overcrowded platforms, full trains, and accumulated delays due to large amounts of people getting on and off the trains.
You can easily find schematic maps publicly-displayed in every Paris station and they are the go-to tool to navigate around the city. As described in detail here these maps follow the schematic map principles of straightening the lines and skewing the scales. This makes the Paris transit map a very convenient, yet imperfect visual of spatial information that sometimes offers a distorted representation of reality.
Believe it or not, I think the map is partly to blame for the horrific traffic conditions: if travelers rely on a distorted map as a single resource to make their route decisions, they tend to choose a path that would seem convenient on the map but is not necessarily be the optimal solution for their chosen route.
Additionally, when commuters make a route decision, they might do so by choosing the path of least resistance, such as choosing a connection with fewer transit stations and vehicle changes. But they are some other intangible factors passengers take into account when making route decisions. These are tricks deciding factors that you can’t find on depicted on a transit map such as hours, fares, waiting times and station availability for instance. If you try to find a balance between all the deciding factors, you’d probably find yourself with an endless variety of possible routes.
The purpose of this research was to put my hypothesis to test and study the actual impact of map design on traffic congestion.
How might we better use transit maps to encourage passengers to use public transportation in the most convenient way?
In this study, I assume that average Joe commuter (or Jacques in our case) would rely more on their own experience rather than transit maps when choosing their routes and tend stick to the same combination of stations due to habit. Considering this assumption and the high number of people commuting every day, the use of schematic maps would fail to reduce traffic congestion. So the idea was to focus on regular commuters and to see if we could change their behaviour if they were shown different maps.
To start my research properly, I needed more assumptions and used my personal experience as an average Jaqueline to define and analyse the two following assumptions.
Assumption 1: People only use a limited number of stations in the entire network
Note: Given the limited resources I had, I decided to perform the research on my own personal traveling behaviour, as a regular Parisian commuter. Of course, it would have been great to access and analyse data from other commuters in a longer time frame. This is a bigger project I’d be happy to work on, so if you’re up for a collaboration, give me a shout-out.
To give you some context, I have been living in Paris for two years now. Slowly but surely I have become more and more familiar with the city and its urban infrastructure as I deal with commuting every day. I spend around two hours per day on public transport. Most of the time, I use the metro and RER (regional express train), but sometimes I opt for buses or tramways. I walk as soon as I get the chance and mostly use my bike during the summertime.
I thought it would be interesting to report my trips each time I travelled around the city. The idea behind this was to determine if there are patterns in my traveling behaviour that could verify if there are some areas that I frequent more often and others where you‘d probably never see me.
So during one month, I reported every single path I took with the sequence of departure stations, transfer stations and arrival stations (D-A pairs). If I was walking or biking, I just reported the closest metro station.
To no one’s surprise, the most common D-A pair was from my home (Château Rouge) to work (Pont de Sèvres), which can be seen by the size of the two biggest ellipses in the diagram above. But interesting enough, my most used pairs required at least one transfer station and an intersection with another metro line. It is interesting to see that those transfer stations are among the most frequented ones from my traveling report.
Therefore, my traveling behaviour depends a lot on the transfer stations and some of my most used stations are stations where I actually never stopped, such as: Trocadéro (line 6 and 9), La Muette (line 9 and RER C) , Charles de Gaulle Etoile (line 2, 6 and 9). This means that the more a station is connected with other lines, the probability for me to go there more frequently is higher.
I wanted to better understand how these stations are connected with each other among the overall network infrastructure and which lead me to my second assumption.
Assumption 2: Some stations are overly full and others are completely empty
To dig deeper into the meaning of my travel report, I decided to get rid of the smaller dots — considering that I only went there occasionally and the probability of me going there again was very low. Ignoring these less frequented stations resulted in an overview of my 20 most visited stations.
As seen in Assumption 1, there are specific focal points related to the network infrastructure and transit lines, which means that traffic is not equally distributed among all the stations in the transit system. I wanted to better understand how these 20 most visited stations interacted within the whole urban infrastructure and especially with the other stations that match my travelling habits (the other stations that lie between all my D-A pair journeys).
In the diagram below you can see the different stations I used according to my traveling report. The top left stations are my most 20 visited stations whereas the other ones in the circle represent the stations I have to go through to reach the final destination. By matching my most visited stations with the other ones on the lines they belong to, a network of connections emerged.
It is interesting to see that some stations have more connections than other ones. And this is due to their position in the network infrastructure as well as the number of lines they belong to. For instance, Gare du Nord (line 4, 5, RER B) is connected with 19 other stations of my network or Chaussée d’Antin La Fayette (line 7, 9) has 31 connections with other stations of my network whereas George V (line 6) only has 4 connections and Kléber (line 6) has 2 connections within my network.
If I only keep the stations that have more than 15 links, this would determine the stations where I am the most likely to go according to my travel habits and the transit infrastructure.
Commuters, especially the regulars Jaques and Jaqueline, choose their path according to their personal travel needs and habits. Following the theory of bounded rationality of Herbert A. Simon (also explored in John Xu’s study case on the impact of the metro map in Washington DC), people make decisions within the limits of the information available to them. This means they make decisions that they think will satisfy their needs without making sure they have considered every single possible option. These decision-makers are seen as “satisficers” and therefore make decisions seeking for personal perceived satisfaction for themselves rather than finding the optimal solution.
When commuters make their route decisions, they choose a path that seems satisfying to them (probably the route that will minimize travel time and cost) without necessarily thinking of what is the most optimal one at that specific time, leading them to ignore important factors that might negatively impact their travel experience.
As there is a clear link between the passenger’s traveling behaviour and the urban infrastructure, the next step would be to collect and cross data from every single commuter. By linking transportation habits with the urban infrastructure network, this could lead to a personalized map that could emphasise the stations each commuter is the most likely to use. By doing this, we could determine the most frequented stations and the less frequented ones and provide an important metric that schematic maps just aren’t able to predict: T-R-A-F-F-I-C!
The big idea is to find out how metro route choices are made by travelers and adjust the map accordingly — in real-time — to help passengers make better-informed decisions about their routes.
Indeed, one single map for everyone would never be enough to depict all the elements that might influence personal travel decisions and meet each passenger’s expectations. However, a personalized route plan, based on subjective travel preferences would be able to meet a passengers’ diversified travel demands. By offering passengers a dynamic path that can be adapted according to real-time crowd information on board, they could make decisions and choose a more comfortable and efficient route. If everyone chooses a more fluid and less crowded route, congestion issues can be reduced and people would spread more homogeneously throughout the entirety of the public transport infrastructure.
Commuting without the stressful hustle and bustle and squishing and squashing? Sounds like a dream to me! If this is something you’re also interested in stay tuned for Part III of this blog series where I’ll be proposing a solution to this little wicked problem…