# Structures Damaged by Wildfire in California 2020-2022: Cal Fire Damage Inspection Images

## Facts

- Provider: Kevin Luo, with data from the California Department of Forestry and Fire Protection (Cal Fire)
- License: CC-BY-4.0 (https://zenodo.org/api/records/8336570)
- Size: 11 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/kevincluo/structure_wildfire_damage_classification
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.8336570
- Category: Environmental Science
- Source: https://zenodo.org/record/8336570
- Page: https://scigantic.com/archive/9707b811-d273-5ee5-abba-8dfeac47afbb
- Keywords: Environmental Science, Parquet, Hugging Face

## About

Over 18,000 field photographs of homes and other structures taken by the California Department of Forestry and Fire Protection (Cal Fire) during damage inspection (DINS) after 18 wildfire events in 2020 to 2022, including the August Complex, Bear, Caldor, Castle, Creek, Dixie, Glass, Mosquito, SCU Complex, Zogg and other fires. Each image carries one of six damage classes from the inspection record: no damage, affected, minor, major, destroyed and inaccessible. The author retrieved the records, published as GIS feature layers on the Cal Fire Hub, processed them into image and tabular form, and republished them with Cal Fire's permission. This Hugging Face copy, uploaded by the dataset author, stores the images and integer labels as nine Parquet shards (about 5 GB) loadable with Hugging Face datasets; the Zenodo record also offers the images as a zip and an additional attribute data archive. Uses include post-wildfire building damage classification, rapid damage assessment and studies of structure vulnerability to fire. Licensed CC BY 4.0 by the author on Zenodo (record 8336570); cite Kevin Luo, 'Image Dataset of Structures Damaged by Wildfire in California 2020-2022'.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("kevincluo/structure_wildfire_damage_classification")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00009-c8a5637a3d1a80e9.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/9707b811-d273-5ee5-abba-8dfeac47afbb
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/9707b811-d273-5ee5-abba-8dfeac47afbb
- Over MCP: `get_schema_card` with id `9707b811-d273-5ee5-abba-8dfeac47afbb` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Environmental Science datasets](https://scigantic.com/topics/environmental-science)
