Outerview/fire-hydrants-dataset

---
language:
- en
pretty_name: Global Fire Hydrants Dataset
license: cc-by-4.0
tags:
- geospatial
- computer-vision
- object-detection
- mapping
- infrastructure
- fire-hydrants
- open-data
- world-models
- mapillary
task_categories:
- image-classification
- object-detection
size_categories:
- 10K<n<100K
annotations_creators:
- machine-generated
language_creators:
- other
multilinguality:
- monolingual
source_datasets:
- original
---

# Global Fire Hydrants Dataset

The **Global Fire Hydrants Dataset** is a geotagged dataset of **14.2K fire hydrants** from around the world, released by **Outerview**, a research lab focused on building world models.

At Outerview, our mission is to **organize the world’s physical information and make it accessible and usable**. This dataset is part of that effort: transforming physical-world infrastructure into structured, machine-readable data for research, modeling, and real-world intelligence.

## Dataset Description

This dataset contains **14,200 examples** of fire hydrants with geographic and image metadata.

Each entry is centered on a single real-world physical feature: a **fire hydrant**.

It is designed for:

- computer vision
- geospatial machine learning
- infrastructure intelligence
- mapping and asset discovery
- multimodal retrieval
- physical-world search and indexing

## Data Source and Processing

The underlying imagery in this dataset is sourced from **Mapillary**.

The dataset was computed, extracted, and structured using the **Outerview API**, which is designed to help index and organize physical-world information at scale.

## Why This Dataset Exists

Most of the world’s physical infrastructure is still difficult for machines to access and reason about.

Fire hydrants are a strong example of this problem: they are common, geographically distributed, visually diverse, and operationally important, yet clean open datasets for them remain limited.

We created this dataset to help make physical infrastructure more searchable, more usable, and easier to work with in machine learning and geospatial systems.

## What’s Included

This release contains **14.2K fire hydrant records** with metadata fields such as:

- unique ID
- latitude
- longitude
- source
- name
- region
- filename

Images are included as part of the dataset release assets.

## Dataset Structure

Typical columns include:

- `id`
- `latitude`
- `longitude`
- `source`
- `name`
- `region`
- `filename`

## Uses

This dataset can be used for:

- fire hydrant classification
- object detection and localization
- infrastructure mapping
- geospatial indexing
- physical-world retrieval systems
- training and evaluating world models
- civic infrastructure research
- real-world AI prototypes

## Coverage

- **Feature type:** Fire hydrants
- **Count:** 14.2K
- **Scope:** Global
- **Modality:** Street-level imagery with geospatial metadata

Coverage is not uniform across all countries or regions. This dataset should be treated as a research and development resource rather than a complete inventory of all fire hydrants worldwide.

## Limitations

- Geographic coverage is uneven
- Image quality and capture conditions vary
- Some hydrants may be partially occluded, distant, blurred, or difficult to identify
- The dataset reflects available imagery and collection coverage, not complete ground truth
- Presence in the dataset does not imply full regional completeness

## Recommended Uses

Recommended for:

- research
- model training
- evaluation
- geospatial analysis
- infrastructure discovery
- retrieval and indexing workflows

Not recommended for:

- safety-critical decisions without independent verification
- emergency response use without validation against authoritative sources
- claims of complete hydrant coverage for any city or country

## About Outerview

**Outerview** is a research lab building world models.

We work on systems that help machines understand **what exists in the physical world, where it is, how it changes, and how to navigate it**.

We believe the physical world should be as searchable and understandable as the digital world. Our goal is to organize the world’s physical information and make it accessible and usable for research, products, and real-world applications.

## Citation

If you use this dataset, please cite:

@dataset{outerview_global_fire_hydrants,
title={Global Fire Hydrants Dataset},
author={Outerview},
year={2026},
publisher={Hugging Face}
}

## License

This dataset is released under the **CC-BY-4.0** license.

## 🔗 API & Full Dataset Access

This is a sample dataset.

The full platform provides:
- Millions of additional locations
- **Timestamps / dates for each capture**
- Access to **billions of real-world images and videos**
- Real-time querying of physical features

Access the full dataset and API:
👉 https://outerview.ai

View API documentation:
👉 https://outerview.ai/developers/docs

## Future Releases

This dataset is part of a broader effort by Outerview to publish structured datasets of physical-world features, infrastructure, and objects for training and evaluating world models.

Future releases will expand into additional categories, geographies, and scales.