---
annotations_creators: []
language:
  - en
language_creators: []
license: cc-by-sa-4.0
multilinguality:
  - monolingual
pretty_name: Street Scene Descriptions (Adaption)
size_categories:
  - 10K<n<100K
source_datasets:
  - extended|https://huggingface.co/datasets/Reubencf/streetview-global
tags:
  - adaption
  - instruction-tuning
  - street-view
  - mapillary
  - geographic
  - vqa
  - image-captioning
  - scene-understanding
  - transportation
  - architecture-design
task_categories:
  - image-to-text
  - visual-question-answering
  - image-classification
task_ids: []
---

![banner](https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/54a29122-2a74-4136-a6b2-d75fa43704e2.png)

This dataset is a remastered version of
[`Reubencf/streetview-global`](https://huggingface.co/datasets/Reubencf/streetview-global)
prepared using [Adaption's](https://adaptionlabs.ai/app/auth) Adaptive Data platform.

# Street Scene Descriptions (Adaption)

10,117 globally-sampled street-level images paired with detailed textual
scene descriptions and rich structured metadata (setting, weather, time of
day, road type, geolocation). Each entry features a natural-language
caption describing the scene along with Adaption-sharpened
`enhanced_prompt` and `enhanced_completion` columns suitable for
instruction tuning.

### Dataset size

10,117 rows. Instruction-tuning format.

### Domain
- Art (64%)
- Architecture-design (30%)
- Transportation (4%)

### Tone
- Descriptive (100%)

## Distributions

### Macro-region coverage (top groupings)

| Region | Count |
|---|---:|
| Europe | 2,245 |
| North America | 1,814 |
| South America | 830 |
| Southeast Asia | 780 |
| East Asia | 663 |
| Africa | 553 |
| Middle East | 442 |
| South Asia | 387 |
| Oceania | 325 |
| Central America | 323 |
| Remote / uninhabited | 285 |

Plus city-level samples from **129 additional `Country – City` buckets.**

### Setting

| Setting | Count |
|---|---:|
| urban | 4,059 |
| suburban | 3,625 |
| rural | 1,429 |
| highway | 785 |
| forest | 132 |
| coastal | 60 |
| desert | 60 |
| mountain | 54 |

### Weather

| Weather | Count |
|---|---:|
| sunny | 6,069 |
| overcast | 3,025 |
| cloudy | 785 |
| rainy | 86 |
| snowy | 45 |
| foggy | 30 |

### Time of day

| Time | Count |
|---|---:|
| day | 9,797 |
| dusk | 199 |
| night | 185 |
| dawn | 23 |

### Road type

| Road type | Count |
|---|---:|
| paved | 9,392 |
| unpaved | 521 |
| cobblestone | 62 |
| dirt | 16 |

## Intended Uses

- Visual question answering on real-world outdoor scenes.
- Image captioning / scene description with geographic diversity.
- Scene classification (setting, weather, time of day, road surface).
- Geo-aware vision-language model evaluation.
- Street-view retrieval and clustering.

## Limitations and Biases

- Sampling is heavier in Europe and North America than in lower-income
  regions, reflecting Mapillary contributor distribution.
- The overwhelming majority of images are daytime and sunny/overcast;
  low-light, night, and adverse-weather scenes are under-represented.
- Scene descriptions, tags, and VQA pairs are model-generated with Google
  Gemma 4 31B (`google/gemma-4-31B-it`) served via Together AI, and may
  contain hallucinations, mislabeling, or biased cultural inferences —
  especially when inferring country from signage or architecture.
  Downstream users should treat them as weak supervision rather than
  ground truth.
- Some images contain incidentally-captured faces, license plates, or
  signage; Mapillary applies automated blurring, but coverage is not
  perfect.
- `region` labels mix coarse macro-regions with fine `Country – City`
  buckets depending on how the sample was collected.

## Source and License

Images are sourced from the public **[Mapillary](https://www.mapillary.com)**
API and are licensed under **CC BY-SA 4.0**, the license under which
Mapillary contributors release their photographs. You must attribute
Mapillary and its contributors, and redistribute any derivative dataset
under the same license.

The annotations (scene descriptions, scene classifications, VQA pairs,
and Adaption-sharpened prompts/completions) in this repository are
released under the same **CC BY-SA 4.0** license.

## Related

- Original corpus (images + structured metadata + VQA):
  [`Reubencf/streetview-global`](https://huggingface.co/datasets/Reubencf/streetview-global)
- Adaption collection:
  [`Reubencf/proper-adaption`](https://huggingface.co/collections/Reubencf/proper-adaption)

## Evaluation Results

- **Quality Gains:**
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/8f682ffd-4dc8-49ef-8989-92df997ca4c0.png" alt="QualityGains" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />

- **Grade Improvement:**
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/bb4a5904-232a-43ed-9b9d-a3e91773df59.png" alt="Grade" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />

- **Percentile Chart:**
<img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/0eb93776-9c5a-4be9-bb26-239db2d20253.png" alt="Percentile Chart" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />

## Citation

If you use this dataset in your research or downstream work, please cite:

```bibtex
@misc{adaption_street_scene_descriptions_2026,
  author       = {Fernandes, Reuben Chagas},
  title        = {Adaption Street Scene Descriptions: StreetView Subset with Enhanced Supervision},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/Reubencf/adaption-street-scene-descriptions}}
}
```

APA-style:

> Reuben Chagas Fernandes (2026). *Adaption Street Scene Descriptions: StreetView Subset with Enhanced Supervision* [Dataset]. Hugging Face. https://huggingface.co/datasets/Reubencf/adaption-street-scene-descriptions