The IAPI (Inclusive Accessibility by Proximity Index) has been tested in Bologna and is now available for application in other testbeds.
By Luigi Carboni, Giovanni Lanza, Paola Pucci, Politecnico di Milano
Following the first trial in Crescenzago (Milan), the Politecnico di Milano team, with the collaboration of TUM, has updated the methodology for constructing and calculating the IAPI and tested it in Bologna to support further development of street experiments in the city. The new version of the IAPI builds on the key outcomes and feedback from the first experiment: while pursuing the same goals of measuring accessibility via active mobility to essential services by differentiating between distinct travel experiences (pedestrians, cyclists, people with reduced mobility), the new version thoroughly updates the methodology and steps for calculation. The new methodology, in synthesis, is based on three subsequent steps:
The first step concerns defining the indicators and characterizing the digital simulation graph.
The list of indicators expanded and consolidated compared to the previous version, identify the presence of conditions that (dis)favour walkability and cyclability concerning the technical performance of the paths (network indicators) and the quality of the surrounding open and built spaces (relational indicators). Indicators of both categories are built by combining open data (OSM), proprietary administrative data and data collected from targeted surveys.
Because of their characteristics, Network indicators are mapped directly onto each arc (the digital reproduction of a neighbourhood’s street) that makes up the digital simulation graph. In contrast, relational indicators are mapped onto a hexagonal grid overlapped with the same graph. This is because relational indicators describe the characteristics of public spaces, forms of use and interactions between open and built spaces and are not directly linked to the technical characteristics and performances of the paths.
The experiences of different users are then simulated by introducing coefficients expressing how the conditions described by each indicator can have a low, medium, or high positive or negative impact on walkability and cyclability based on the user's ability and possibility to move. The coefficients are calibrated for each network and relational indicator. While the relational indicators coefficients are considered in the calculation later (step 3), the network indicators coefficients are calculated directly in the first step of the methodology. Their effect is expressed as a change in speed along the arcs of the network, which increases compared to the average value of each profile of the user under conditions favouring walkability and cyclability, and vice-versa. Such an approach allows calculating a different speed for each arc of the network based on the current characteristics (dis)favouring active mobility for the three profiles. In addition, this calculation makes it possible to develop an interesting intermediate outcome: a map that ranks urban streets according to the extent to which current technical performances do or do not promote walkability and cyclability, thus offering a useful assessment tool for planning activities of public administrations.
The second step involves the mapping of services to be considered essential and included in the accessibility assessment.
This mapping is done by considering each facility’s location, performance, and hourly availability. Once mapped, the services are included in the same simulation base characterized in the previous step.
The third step concerns the final IAPI calculation. It starts by calculating isochrones on the graph characterized through network coefficients by considering three-time thresholds (5, 10, and 15 minutes).
The speeds assumed in the simulation are thus different for each arc of the road network based on the characterization performed in step 1, returning a more detailed and diverse geography of each service's catchment area compared to an approach assuming constant speeds within the network. Calculation of isochrones is performed using, as starting points, each service, as mapped in step 2.
Subsequently, the regular hexagonal used for mapping the relational indicators is overlapped with the isochrones, which are sampled assigning different scores to each hexagon based on their accessibility values (higher for lower time thresholds and vice versa). This value is then multiplied by the coefficients of the relational indicators insisting on each hexagon.
Accessibility levels to primary and lower secondary schools in Bologna (kindergartens, primary and middle schools) calculated using the IAPI for pedestrians (left) and people with reduced mobility (right). Note that indicators 3.2 and 3.3 related to public perceptions were applied only in the Bolognina area (bordered in black) due to the experimental collection of crowdsourced data limited to this part of the city. Ⓒ Politecnico di Milano
Finally, the values of each hexagon could be summed n times as many as services considered in the basket and then normalized to allow comparison between different areas in the neighbourhood. Due to the cumulative principle, the result is a composite indicator in which higher values correspond to high levels of accessibility understood both in physical and relational terms. In addition to the updates provided on the methodological structure, other possibilities for improving the index were experimented with. The most relevant concern:
The use of crowdsourced data for calculation in the case of Bologna as an essential step to make the tool more sensitive to the socio-spatial context of application. The data used, collected through a Commonplace-based survey, allowed 1) to obtain baseline data not available in open format to map specific relational indicators describing inhabitants' perceptions toward public spaces of a neighbourhood 2) to weigh The coefficients for The different profiles based on stated preferences collected through the survey, asking respondents to indicate which walkability/cyclability factors were most relevant to them.
The success of this first experiment prompted Polimi and FIU to promote a newly updated questionnaire, recently released, open to all Bologna citizens. The questionnaire will also investigate the relevance of essential services to introduce priorities and preferences concerning different respondent profiles.
The development of an automated calculation model in a GIS environment to facilitate index calculations and allow a high level of customization and transferability to different contexts
The draft of Web GIS dashboards for visualizing the IAPI as an intuitive tool for dissemination and use by public administrations, policymakers, and planning practitioners