comarti15/accessibility-rotterdam-the-netherlands

Dataset Card for Mobility and Accessibility in Rotterdam's Outdoor, Public Spaces

<!-- Provide a quick summary of the dataset. -->

Compilation of 1,883 street-level photos of Rotterdam, The Netherlands in spring 2025. Labeled as either accessible or inaccessible, in relation to mobility friendliness for people with wheelchairs, strollers, walkers, etc.

Dataset Details

The dataset was created as part of a project for a course called Unboxing the Algorithms (Course ID: CM4606) at the Erasmus School of History, Culture, and Communication (ESHCC), supervised by Dr. JF Ferreira Gonçalves. The primary goal of the project was to assist the municipality of Rotterdam in identifying barriers to mobility to ultimately promote inclusivity in public spaces.

Dataset Description

<!-- Provide a longer summary of what this dataset is. -->

This dataset contains 1,883 images depicting outdoor, public, and pedestrian spaces in Rotterdam, The Netherlands. The images were sourced from Google Street View and other online platforms, as well as through manual photography. The goal of this dataset is to train machine learning models that can automatically detect accessibility barriers for people with mobility impairments. The dataset was manually labeled to identify features such as blocked sidewalks, inaccessible ramps, and other physical obstacles in public spaces.

• —Curated by: Cora Martin, Yu Vos, Agnė Piepaliūtė, Lucas Holef
• —License: apache-2.0

Uses

<!-- Address questions around how the dataset is intended to be used. -->

Direct Use

<!-- This section describes suitable use cases for the dataset. -->

The dataset is intended to be used for training and evaluating computer vision models to detect accessibility barriers in outdoor public spaces. Researchers and urban planners could use these models to map inaccessible areas and promote more inclusive infrastructure.

Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->

The dataset is not suitable for purposes unrelated to accessibility analysis. Misuse includes using this datat to generate discriminatory or biased results.

Dataset Structure

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->

The dataset consists of:

• —Images: 1,883 images of outdoor public spaces in Rotterdam, The Netherlands.
• —Labels: Binary labels indicating the presence or absence of accessibility barriers.
• —Metadata: Image filenames.

Dataset Creation

Curation Rationale

<!-- Motivation for the creation of this dataset. -->

The dataset was created to support the development of machine learning models capable of identifying physical accessibility barriers in public spaces in Rotterdam, The Netherlands. This project was motivated by the lack of large-scale datasets for accessibility analysis in urban environments, particularly in Rotterdam.

Source Data

<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->

Data Collection and Processing

<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->

The primary source of data was Google Street View, accessed through the Google Maps API. The research team used a Python script to generate 5,000 random geographic points within predefined latitude and longitude boundaries in Rotterdam. The script then fetched a 640x640 pixel image from each point. After removing null or repeated images, 3,579 images remained.

Some images were manually taken by the researchers using smartphones, and some were sourced from public platforms such as Google Images and Bing. Images that did not depict outdoor, walkable public spaces were manually removed.

Who are the source data producers?

<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->

The source data producers include Google (via Google Street View), online image platforms, and the researchers themselves who manually captured some images in Rotterdam.

Annotations

<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->

Annotation process

<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->

The labeling process was carried out manually by four researchers (Cora Martin, Yu Vos, Agnė Piepaliūtė, Lucas Holef). Each researcher was assigned approximately 900 images to annotate. Images were labeled based on the presence of accessibility barriers, such as:

• —Blocked or non-existent sidewalks
• —Absence of ramps
• —Improperly parked electric scooters or bicycles
• —Lack of lowered pavements

Who are the annotators?

<!-- This section describes the people or systems who created the annotations. -->

The annotators were the four researchers listed as dataset curators. They are students of the Unboxing the Algorithms course (Course ID: CM4606) at the Erasmus School of History, Culture, and Communication. No external annotators were hired.

Personal and Sensitive Information

<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->

The dataset does not contain any personally identifiable information. While some images may incidentally contain people or vehicles, their faces are not visible or are blurred out.

Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

This dataset has several limitations:

• —Geographical bias: The dataset only covers the city of Rotterdam and may not generalize to other cities or countries.
• —Subjective labeling: Despite the use of standardized labeling guidelines, the perception of what constitutes an accessibility barrier may vary between annotators.
• —Static images: The images represent a snapshot in time and may not reflect recent or future changes in infrastructure.

Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users should be made aware of the geographical and subjective bias present in the dataset. Additionally, users should avoid deploying models trained on this dataset in regions with vastly different urban infrastructure.

Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

BibTeX:

@misc{mobilityaccessibilityrotterdam, title={Mobility and Accessibility in Rotterdam's Outdoor, Public Spaces}, author={Cora Martin and Yu Vos and Agnė Piepaliūtė and Lucas Holef}, year={2025}, note={Erasmus University Rotterdam} }

APA:

Martin, C., Vos, Y., Piepaliūtė, A., & Holef, L. (2025). Mobility and Accessibility in Rotterdam's Outdoor, Public Spaces. Erasmus University Rotterdam.