D4AMS handbook: How and why to use and lessons learned

Street Experiments

Street experiments, a specific form of transition experiments, are “intentional and temporary changes in the use, regulation and form of the street, with a shift from motorized to non-motorized dominance and aimed at exploring systemic changes in urban mobility and public life” (Bertolini, 2020; VanHoose et al. 2022). They are becoming increasingly popular in cities worldwide, particularly during and after the COVID-19 crisis (Verhulst et al., 2023). Occurring in many different shapes and forms, these interventions show benefits in terms of more local retailing (Hass-Clau, 1993), increased health (Kingham et al., 2020; Mueller et al., 2020; Wolf et al., 2014), increased safety (Lee and Kim, 2021), improved social connections with neighbours (Appleyard, 2021; Kingham et al., 2020; Pandit et al., 2021), and reduced pollution (Brimblecombe and Lai, 2020). However, next to these socio-economic and cultural aspects and more relevant to the scope of this work package, they show huge potential in enabling systemic change in urban mobility (Bertolini, 2021; VanHoose et al., 2022; Verhulst, 2023). Yet, it remains difficult to assess the impacts of street experiments on mobility. Effects on local traffic flows, overall mobility, and travel behaviour remain somehow ambiguous (Melia and Calvert, 2023).

For example, the closure of a street to private vehicles might produce unintended consequences on nearby streets and even the entire neighbourhood. This notion is frequently cited by critics of such initiatives (Henderson, 2013; Goodman et al., 2021). An illustration can be seen in Ghent, where the implementation of a new traffic plan in 2017 led to a substantial shift away from private motorized vehicles citywide (City of Gent and TML, 2019). While in many streets a transition from cars to active modes was observed, certain streets, like Rozemarijnstraat, saw an increase in car traffic. These are what policymakers refer to as “Loser streets”, a concept coined by Appleyard (1980). This remains to pose challenge in managing adverse effects to policy makers and mobility experts.

In terms of traffic impacts, the literature generally outlines three main patterns of traffic displacement when streets are temporarily or permanently closed off to car traffic. Cairns et al. (1998) formulated three hypotheses to describe this phenomenon. The first hypothesis suggests that such measures won’t reduce traffic at all. The second proposes that traffic will redirect or adjust its timing. Melia and Calvert (2023) discovered that the specific outcomes vary depending on where the street closure occurs. While a pedestrian scheme at a small scale resulted in longer car trips, a temporary bridge closure led to traffic diversion and some traffic evaporation. This last observation aligns with the third hypothesis of Cairns et al. (1998), which describes the phenomenon when a certain number of activities and trips no longer take place, causing this traffic to disappear or evaporate (Cairns et al., 2002; Hass-Clau, 1993).

The actual magnitude of traffic evaporation is typically modest, but when combined with a shift toward sustainable transportation modes and redirection toward alternative main roads, it can yield overall benefits. For instance, Barcelona’s superblock program showed a minor traffic evaporation effect alongside a slight increase in traffic on surrounding streets within the superblocks (+2%) (Nello-Deakin, 2022). This finding echoes observations from low-traffic neighbourhoods (LTNs) in London, where traffic decreased within the LTNs while a slight rise was noted on boundary roads (Thomas and Aldred, 2024).


Presently, there is a shift in urban planning from a focus on cars to prioritizing people. Cities are exploring alternative uses of public space and reconsidering mobility strategies (Bertolini, 2020; Bibri et al., 2020; Smeds, 2023; VanHoose et al, 2022). This move towards a ‘post-car’ city aims at creating more liveable urban environments, advancing climate neutrality, and addressing various other objectives related to health, education, social well-being, and economic advantages (Jacobs, 1961). Notably, there is a growing emphasis on initiating a modal shift to replace cars with more active and sustainable transportation modes, particularly in densely populated urban areas. This shift is mirrored in transport policy development regarding parking spaces and the integration of new mobility trends like shared and micro-mobility options. Additionally, there’s a momentum towards modifying street design and functionality to optimize available space. This goes along with the public, academic and political discussion on how public space should be designed, which functions it should accommodate and how car traffic can still fit into this. Examples include strategies such as circulation plans, 15-minute cities, superblocks, tactical urbanism, and street experiments (Bertolini, 2020; Lydon and Garcia, 2015; Moreno et al., 2021; Mueller et al., 2020; Smeds, 2023). These concepts gained prominence during the COVID-19 pandemic, aiming to catalyse systemic change and allocate more space to citizens and active transportation modes (Pandit et al., 2021; Thomas et al., 2022; Noi et al., 2022; Verhulst et al., 2023). Overall, these initiatives can be coined as transition experiments (Nevens et al., 2013; Roorda et al., 2014) in city streets, which can be defined as “intentional, temporary changes in street use, regulation and/or form, aimed at exploring systemic change towards a ‘post-car’ city”.

In this study, which is incorporated into work package 3 of the EX-TRA project, we aim to identify opportunities and threats of new alternative mobility options for attractive city streets and accessible city districts. Concretely, this work package focusses on the mobility impacts of two distinct, but often conjoining, strategies that are associated with a transition to sustainable mobility: street experiments and shared-mobility solutions. It will do so by simulating hypothetical scenarios through an agent-based model approach for four different case studies: Ghent, Bologna, Amsterdam and Munich.

Shared mobility solutions

A prevalent critique of car traffic calming schemes revolves around the absence of viable alternatives for residents to using their personal vehicles (Henderson, 2013). However, remarkably, initiatives such as circulation plans, car-free neighbourhoods, low-traffic neighbourhoods (LTNs) and “Living Streets” increasingly coincide with the introduction or expansion of shared mobility options (Bertolini, 2020). Over the last decade, various forms of shared- and micro-mobility have been introduced to cities worldwide. These alternative transportation modes offer users the flexibility of using vehicles for specific durations without the need for private ownership, presenting various benefits such as reduced demand for parking space and increased accessibility. For instance, they tend to be a likeable substitute for walking for last and first mile trips to public transport (Shaheen & Chan, 2016). However, some cities face challenges associated with these options, including congestion of micro-mobility on pedestrian pathways, environmental impacts, and unwanted shifts in transportation modes, such as a transition from public transport to car-sharing (Casier et al., 2021; Martin and Shaheen, 2011).

While the costs and benefits of shared transportation modes are well-documented, research on the combined impact of street closures and shared mobility options on mobility is scarce (Guyader et al., 2021; Machado et al., 2018). For example, the emergence of shared mobility options has provided an opportunity to repurpose parking spaces for alternative activities like parklets, contributing to improvements in liveability. Generally, it could be assumed that car-sharing reduces the demand for parking space while street closures decrease the supply of available space for cars (Glaser and Krizek, 2021; Tchervenkov et al., 2018). However, a more comprehensive understanding of both the advantages and disadvantages of such interventions and their potential positive effects on shared mobility options is necessary.

Objectives and research questions

The lack of insight into the combined impacts of street experiments and shared mobility solutions on urban mobility forms the starting point for formulating our research questions.

The research questions can be divided into two sub questions:

  1. What is the impact of street closures on travel behaviour, in terms of
  • modal split shifts
  • car traffic volumes
  1. What is the impact of street closures combined with shared mobility solutions on travel behaviour, in terms of
  • modal split shifts
  • car traffic volumes

The study outcomes will then be translated into policy guidelines directed at public authorities and mobility practitioners. As a final product, the study outcomes and the policy guidelines will be integrated in an online dashboard, called Dashboard for Alternative Mobility Scenarios (D4AMS) that will be open to the public and aims at disseminating scientific knowledge to public authorities and mobility practitioners in order to assist them in developing effective local policy measures.

The rest of this report is divided into four sections. The initial part delves into the methodologies employed. Following that, the second section will comprehensively discuss the findings of each case study. The third and principal section articulates the primary policy guidelines gleaned from the study. Lastly, the fourth section will provide the study’s conclusion.


This section delves deeper into the methodologies applied for this research. It consists of two parts. The first part addresses the modelling approach, the second part explains the qualitative workshop approach.

Computational Methodology development: Agent-based Modelling using MATSim

Before delving into the specifics of the MATSim methodology, it’s important to establish some foundational concepts about agent-based modelling (ABM). ABM is a computational modelling approach centred around autonomous entities known as agents. These agents represent individual actors whose behaviours contribute to emergent patterns at a larger scale. In our research, we employ multi-agent simulation to analyse collective behaviour. This method operates on a micro-scale level, where agents are endowed with rules dictating permissible actions based on their personal objectives. For instance, in a transportation context, agents may be restricted to certain modes of travel or allowed access to specific streets. Their objectives could range from reaching work punctually to dining at particular restaurants.

ABM enables us to implement these general rules and individual objectives across all agents. By utilizing Monte Carlo methods, we strive to strike a balance wherein each agent optimizes its utility within the defined rule set, thus facilitating the achievement of their respective goals. Each study in MATSim always follows the same loop (cf. Figure 1: MATSim loop). This loop is similar to what all agent-based models try to achieve, namely an optimal state of being for all agents. To start the MATSim loop, you need to provide the model with input data in the correct form and with the correct description. Next starts a mobility simulation (mobsim) for one day in the life of an agent, in this step all agents execute their daily schedule within the given set of rules. After the mobsim, each agent gets a utility score for his day. This score is a quantification of an agent’s time utility and is based on multiple elements such as arrival time at an activity, time spent in a car, on a bicycle, work, etc. (Axhausen et al., 2016; Balać & Hörl, 2021; Ciari et al., 2016; Horni et al., 2009)

Once each agent has received a score, the ‘replanning’ phase can begin. All agents try to optimize their personal score/utility by changing their personal daily plans. E.g. change their route, leave for work earlier or use a different mode to go to another location. After changing their daily plans, the loop starts again with a mobsim. We can ask MATSim to repeat this loop multiple times until a pareto equilibrium is reached and no more agents can improve their score. Once this state is reached we have obtained the optimal daily plans for the agents and the analyses can begin (Axhausen et al., 2016).

It is important to note that we can change the MATSim framework in several ways. One way is to change the network, i.e. adding or removing nodes and links or making a street car-free. We may also be able to add new facilities related to street experiments, e.g. recreation spots or a pop-up bar in a specific street. Another way is to change the rules that agents have to follow, e.g. adding the option of car sharing at certain locations (Balać et al., 2019; Balać & Hörl, 2021; Ciari & Balać, 2016; Horni et al., 2009). A third option is to change the way the utility/score is calculated. Normally, this is calculated based on agents’ time expenditure, but we could change this to see what the effect is when agents try to optimize, for example, their distances travelled or their CO2 emissions.

As an end product, MATSim always provides us with the optimal daily plans of the agents, within the given context of the model. These can be used to analyse the accessibility of facilities or modes (Axhausen et al., 2016; Ziemke et al., 2016). Furthermore, it allows us to see the effects, such as which roads will be congested, how many shared cars are used, in which locations shared cars are used most, at what time of day they are used, at what distances they are used, etc. (Ciari & Balać, 2016).

Model assumptions and limitations

Multiple assumptions are in place when using MATSim, as already discussed above. The main reason for this is simplifying reality and turning it into a clear, understandable framework that facilitates analysis, decision-making, and the exploration of potential scenarios within the simulated environment. Below we list some of the main assumptions.

  • Agent Rationality: MATSim assumes that agents behave rationally, meaning they make decisions that maximize their individual utility or satisfaction given the constraints and information available to them.
  • Utility Maximization: Agents in MATSim seek to maximize their utility by selecting actions that best align with their preferences and objectives, such as minimizing travel time or cost.
  • Discrete Choice Modelling: MATSim assumes that individuals make discrete choices among a finite set of alternatives. For example, individuals may choose between different transportation modes or routes based on their preferences and constraints.
  • Dynamic Environment: MATSim assumes that the environment is dynamic, with changes occurring over time due to agent interactions, infrastructure modifications, and external factors such as weather and traffic conditions.
  • Feedback Loop: MATSim operates in an iterative manner, with agents making decisions based on the current state of the system and then influencing future states through their actions. This feedback loop allows for the simulation of complex, dynamic systems.
  • Homogeneous Agents: While MATSim can accommodate heterogeneous agents with different characteristics and behaviours, it often assumes homogeneity among agents within a particular population or group, simplifying the modelling process.
  • Perfect Information: MATSim assumes that agents have perfect information about the state of the system, including travel times, costs, and available alternatives. While this assumption may not always hold in the real world, it simplifies the modelling process and allows for the simulation of agent behaviour under ideal conditions.
  • MATSim only develops ‘one day’ scenarios, and consequently only simulates the mobility impacts of these type of interventions for a specific day. This means that long-term impacts are not examined, for example, if the implementation of street experiments could enable systemic change in urban mobility (Bertolini, 2020; VanHoose & Bertolini, 2023).

These assumptions help shape the structure and functioning of MATSim models, allowing researchers and policymakers to simulate and analyse various transportation scenarios and policy interventions.

Scenario building

Scenario building in MATSim involves creating detailed representations of real-world transportation systems. It includes gathering various data to construct a simulation environment. This process starts with collecting and organizing relevant data, then building models that capture how people move around (their activities and travel patterns). With these models, analysts can simulate individual behaviours within the system. Scenario building allows users to test different policies, changes in infrastructure, or urban development scenarios to see how they might affect travel behaviour, congestion, and the environment. MATSim’s flexibility lets researchers and planners create scenarios suited to their specific needs, aiding in transportation planning and policy decisions.

This allows to tackle a wide range of different policy decisions and problems. Examples of this are the simulation of the effects road pricing, infrastructure alterations (e.g. removing or adding a bridge), noise effects, accidents, BRT, new public transport lines, and more.

Read the full report here

Ex-plore more


Developing Strategies for Change During Street Experiments

Many factors at play when it comes to organizing successful street experiments. In our research, we identified enablers and barriers that either support or limit street experiments on their way to triggering system change. Additionally, we recognized the role that municipalities adopt towards street experiments as particularly important in this process. The Developing Strategies for Change workshop brings these findings together and offers an interactive format for civil servants to explore different roles and enable collaboration between departments.

Show more…

Un indice per misurare la accessibilità di prossimità

The Inclusive Accessibility by Proximity Index (IAPI) measures accessibility to essential services using GIS, focusing on conditions that enhance or hinder walkability, cyclability, and social interactions at the neighbourhood level. Its implementation in Bologna allowed for mapping the quality of pedestrian and cycling routes, evaluating accessibility to neighbourhood services via active mobility, and assessing the impact of pedestrianization interventions on the quality of routes and public spaces. With its ease of calculation, transferable approach using open-source data, and the ability to update indicators and coefficients, IAPI can support the development of multi-sector policies at various scales.

Article published in Italian.

Show more…

Measuring accessibility by proximity for an inclusive city

Accessibility is crucial for social inclusion, influenced by transport systems, land use, temporal availability, and individual features. It measures people’s ability to engage in social life and activities contributing to their well-being. This paper introduces the Inclusive Accessibility by Proximity Index (IAPI), designed to assess accessibility to essential services and activities for local residents. IAPI considers the physical and perceptual characteristics of urban spaces and paths, reflecting different mobility needs and habits. It guides urban planning to promote walkability, cyclability, and active mobility, aiming for a sustainable and inclusive city. Using Bologna, Italy, as a testbed, the paper details the IAPI methodology, results, and steps for scalability and context sensitivity.

Show more…

Making walking irresistible: enabling level-of-service measures to achieve their potential

Despite walking’s exceptional benefits, it receives surprisingly little attention. To address this, the potential of “level of service” (LOS) measures to highlight the status of walking is investigated. A literature survey on various LOS measures reveals their distinct evolutionary paths and the lack of true commensurability across modes. A micro-simulation modelling exercise suggests pedestrians fare worse than drivers, even where walking is promoted, confirming measurement anomalies across modes. The availability of “ideal speeds” is crucial for commensurability; thus, a critical assessment of “free-flow” speeds for vehicles proposes using sprinting speeds for pedestrians to gauge performance.

Show more…

The value of street experiments for mobility and public life: Citizens’ perspectives from three European cities

‘Street experiments’ (SE) temporarily reallocate street space from traffic to people. Existing research often evaluates SE from upscaling or public acceptability perspectives, but few studies explore what citizens value about SE in everyday life. This paper addresses this gap by analyzing how 458 citizens value five SE parklets and plazas in London, Munich, and Bologna. Using an inductive, qualitative survey method, it identifies valued mobility and public life dimensions, considering use value and social meanings. The findings show citizens prioritize SE’s public life benefits—enhanced streetscape, public space availability, and social interaction—over active mobility benefits. This framework aids understanding citizen evaluations and informs better SE design.

Show more…

Urban transport experimentation: a network or hybrid governance process?

This chapter examines the governance of urban transport experimentation, emphasizing the role of collaboration and partnerships beyond government for policy innovation. It explores two propositions: whether municipal governments should primarily enable activities by civil society and private actors, and if voluntary experiment partnerships with shared sustainability goals effectively co-create policy innovations. Using empirical research from 108 transport experiments in Bristol and New York City (1996-2016), it finds that non-state involvement and network partnerships can be crucial but are not essential. Instead, urban transport experimentation is better characterized as ‘hybrid governance,’ featuring multiple co-existing governance modes.

Show more…

StreetEcho toolkit: a how-to guide

StreetECHO Toolkit Guidance provides an overview of the tool and guidance for the engagement of citizens in street transformation. It initially explores concepts underpinning the development of the toolkit, particularly examining streets as convivial public spaces and the imperative of bringing citizen perspectives and knowledge into discussions on street planning and design processes. In emphasising the need for participatory processes, the report highlights the development of accessible digital tools as crucial pathways towards achieving these goals. After reviewing a series of such existing tools, the document introduces the novel StreetECHO toolkit, highlighting its added value and innovative features, and instructing readers on how it’s utilised.

Show more…

Assessing experiments: Citizens perceptions and values

This report presents results from a study that examines citizens’ perceptions and values regarding street experiments in three European cities: London, Munich, and Bologna. The aim of the research was to understand how citizens value different dimensions of street experiments and to provide insights into the design of effective and inclusive street interventions. The study posits that by understanding and incorporating citizens’ values and preferences, practitioners can design street experiment initiatives that are more effective, inclusive, and aligned with the needs and aspirations of the community.

Show more…

How Does Pedestrian Accessibility Vary for Different People? Development of a Perceived User-Specific Accessibility Measure for Walking (Paws)

Current accessibility measures often overlook the diverse needs of different user groups, leading to a mismatch between calculated and perceived accessibility. This paper proposes a new method that accounts for individual perceptions and walkability needs, developing Perceived user-specific Accessibility measures for Walking (PAWs) for seniors, children, women, and wheelchair users. By adjusting the Geo Open Accessibility Tool (GOAT) and using the Analytic Hierarchy Process (AHP), the most important walkability attributes are incorporated and weighted. Results from Munich reveal a nuanced understanding of pedestrian infrastructure, aiding urban planners in creating more inclusive, equitable environments that enhance quality of life and access to amenities.

Show more…