StreetReview Dataset

Overview

StreetReview is a curated dataset designed to evaluate the inclusivity, accessibility, aesthetics, and practicality of urban streetscapes, particularly in a multicultural city context. Focused on Montréal, Canada, the dataset combines diverse demographic evaluations with rich metadata and street-view imagery. It aims to advance research in urban planning, public space design, and machine learning applications for creating inclusive and user-friendly urban environments.

Overview
Dataset Structure
Methodology
- Participatory Evaluation Process
- Data Collection
Data Fields
- Metadata
- Evaluations
Usage
- Cloning the Repository
- Example Code
License
Citing StreetReview
Contributing
Contact

Dataset Structure

The StreetReview dataset is organized as follows:

Root Directory

metadata.csv: Comprehensive metadata for each evaluation point.
street_eval/: CSV files containing evaluation data for individual street sections.
street_img/: Street-view images categorized by street and section.

Street Image Data

Images are stored in street_img/ and organized into folders by street and section, with three perspectives per section (_main, _head, _tail). Example structure:

street_img/
├── i01_cote_sainte_catherine_main/
│   ├── main_001.jpg
│   ├── main_002.jpg
│   ...
└── i02_rue_berri_main/
    ├── main_001.jpg
    ├── main_002.jpg
    ...

Street Evaluation Data

Evaluation data is stored in street_eval/ as CSV files named after their corresponding street section. Example:

street_eval/
├── i01_evaluations.csv
├── i02_evaluations.csv
...

Methodology

Participatory Evaluation Process

The dataset was created using a participatory approach to capture diverse urban experiences:

Individual Evaluation: Participants rated 20 street on four criteria using a color-coded system.
Group Evaluation: In focus groups, participants reassessed images collectively and refined their evaluations.

Data Collection

Participants: 28 individuals contributed to criteria development; 12 participated in detailed evaluations.
Evaluation Points: 60 points across 20 streets, with two images per point.
Dataset Expansion: Up to 250 images per point, rotated for diversity.

Data Fields

Metadata

The metadata.csv file contains attributes such as:

Field	Description
`point_id`	Unique identifier
`sidewalk_width`	Width of sidewalks
`greenery_presence`	Presence of greenery
`building_height`	Height of adjacent buildings
...	...

Evaluations

Each CSV file in street_eval/ includes ratings from various demographic groups:

Field	Description
`lgbtqia2+_accessibility`	Accessibility rating by LGBTQIA2+
`elderly_male_practicality`	Practicality rating by elderly males
`group_inclusivity`	Inclusivity rating by groups of 3-5 diverse individuals
...	...

Usage

Cloning the Repository

Clone the repository with:

git clone https://huggingface.co/datasets/rsdmu/streetreview

Example Code

import pandas as pd
from PIL import Image
import os

# Load metadata
metadata = pd.read_csv('metadata.csv')

# Load evaluation data
eval_data = pd.read_csv('street_eval/i01_evaluations.csv')

# Display an image
image_path = 'street_img/i01_cote_sainte_catherine_main/main_001.jpg'
image = Image.open(image_path)
image.show()

License

Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

Citing StreetReview

@dataset{streetreview2024,
  title = {StreetReview Dataset: Evaluating Urban Streetscapes for Inclusivity and Accessibility},
  author = {RSDMU},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/rsdmu/streetreview}
}

Contributing

We welcome contributions! Please fork the repository, make changes, and submit a pull request.

Contact

For inquiries, contact:

Email: Rashid Mushkani
Website: Rashid Mushkani
GitHub: RSDMU

Datasets:

rsdmu
/

streetreview