README.md

---
annotations_creators:
- no-annotation
language_creators:
- found
language:
- en
license:
- cc-by-4.0
multilinguality:
- monolingual
size_categories:
- 1K<n<10K
source_datasets:
- original
task_categories:
- image-classification
- object-detection
pretty_name: GraptoloideaSpecimensDataset
tags:
- graptoloidea
- paleontology
- specimens
- fossils
- biology
- earth-science
dataset_info:
  features:
  - name: Suborder
    dtype: string
  - name: Infraorder
    dtype: string
  - name: Family (Subfamily)
    dtype: string
  - name: Genus
    dtype: string
  - name: tagged species name
    dtype: string
  - name: image
    dtype: string
  - name: Stage
    dtype: string
  - name: mean age value
    dtype: float64
  - name: Locality (Longitude, Latitude, Horizon)
    dtype: string
  - name: Reference (specimens firstly published)
    dtype: string
  - name: image_local_path
    dtype: 'null'
  splits:
  - name: train
    num_bytes: 587726
    num_examples: 977
  - name: test
    num_bytes: 124910
    num_examples: 209
  - name: validation
    num_bytes: 126456
    num_examples: 211
  download_size: 88922
  dataset_size: 839092
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
  - split: validation
    path: data/validation-*
---
# Dataset Card for Graptoloidea Specimens Imaging

## Table of Contents
- [Table of Contents](#table-of-contents)
- [Dataset Summary](#dataset-summary)
- [Dataset Preprocessing](#dataset-preprocessing)
- [Dataset Description](#dataset-description)
  - [Supported Tasks](#supported-tasks)
  - [Languages](#languages)
- [Dataset Structure](#dataset-structure)
  - [Data Instance](#data-instance)
  - [Data Fields](#data-fields)
  - [Data Splits](#data-splits)
- [Dataset Creation](#dataset-creation)
  - [Curation Rationale](#curation-rationale)
  - [Source Data](#source-data)
  - [Data Processing](#data-processing)
- [Bias, Risks, and Limitations](#bias-risks-and-limitations)
- [Citation](#citation)


## Dataset Summary
This dataset offers a detailed examination of Graptoloidea specimens, featuring attributes like image file paths, suborder, infraorder, family (including subfamily), tagged species names, geological stages, mean age values, and locality details (with coordinates and horizon information), complemented by original reference citations for each specimen. It serves as a comprehensive resource for paleontological research, emphasizing morphological and stratigraphic analysis of these ancient colonial animals.

## Dataset Preprocessing
This dataset doesn't download the images locally by default. Instead, it exposes URLs to the images. To fetch the images, use the following code (Make sure in the correct environment):

```python
from concurrent.futures import ThreadPoolExecutor
from functools import partial
import io
import urllib
import PIL.Image
from datasets import load_dataset
from datasets.utils.file_utils import get_datasets_user_agent
USER_AGENT = get_datasets_user_agent()
def fetch_single_image(image_url, timeout=None, retries=0):
    for _ in range(retries + 1):
        try:
            request = urllib.request.Request(
                image_url,
                data=None,
                headers={"user-agent": USER_AGENT},
            )
            with urllib.request.urlopen(request, timeout=timeout) as req:
                image = PIL.Image.open(io.BytesIO(req.read()))
            break
        except Exception:
            image = None
    return image
def fetch_images(batch, num_threads, timeout=None, retries=0):
    fetch_single_image_with_args = partial(fetch_single_image, timeout=timeout, retries=retries)
    with ThreadPoolExecutor(max_workers=num_threads) as executor:
        batch["image"] = list(executor.map(fetch_single_image_with_args, batch["image"]))
    return batch
num_threads = 20
dset = load_dataset('Graptolodiea-Speciemens-Imaging.py')
dset = dset.map(fetch_images, batched=True, batch_size=100, fn_kwargs={"num_threads": num_threads})
```
## Dataset description


### Supported Tasks
- **Paleontological Analysis and Classification**: Utilizing the dataset for detailed classification of Graptoloidea species, including sorting by suborder, infraorder, and family. Fundamental for researchers in paleontology to understand evolutionary trends and species distinctions.
- **Age Estimation and Stratigraphic Correlation**: Leveraging mean age values and stage data to estimate the geological age of specimens and correlate them with stratigraphic layers. Crucial for geologists and paleontologists in mapping the geological timeline and understanding the Earth's history.
- **Geographical Distribution Study**: Analyzing locality data to study the geographical distribution and migration patterns of Graptoloidea species. Can reveal insights into ancient ecological conditions and biogeographic events.
- **Morphological Analysis**: Using the provided specimen images for morphological studies, enabling the identification of unique features and variations within the Graptoloidea order. Important for taxonomic classification and evolutionary studies.
- **Data-Driven Paleobiology**: Applying machine learning and statistical methods to uncover patterns and relationships in Graptoloidea evolution, diversity, and extinction events.
- **Morphometric Analysis**: Employing image processing techniques to measure and analyze morphological features of the specimens, such as length, shape, branching patterns, and other key characteristics.
- **Virtual Reconstruction and 3D Modeling**: Using specimen images to create detailed 3D models of Graptoloidea for virtual reality experiences, aiding in both research and educational endeavors.
- **Educational and Outreach Tools**: Developing interactive tools and applications for educational purposes, using specimen images to engage and teach students and the public about Graptoloidea and paleontology.
- **Crowdsourcing and Citizen Science Projects**: Allowing citizen scientists to access and annotate the images, contributing to data collection and analysis efforts.

### Language
- **English**

## Dataset Structure

### Data instance
```
{
  'Suborder': 'Axonophora Frech, 1897',
  'Infraorder': 'Neograptina Štorch & others, 2011',
  'Family (Subfamily)': 'Dimorphograptidae Elles & Wood, 1908 (no subfamily)',
  'Genus': 'Akidograptus',
  'tagged species name': 'Akidograptus ascensus',
  'image': 'https://raw.githubusercontent.com/LeoZhangzaolin/photos/main/14545Akidograptus_ascensus.jpg',
  'Stage': 'Rhuddanian, Llandovery (early Silurian)',
  'mean age value': 442.3,
  'Locality (Longitude, Latitude, Horizon)': 'Huangshu Village in Anji County, Zhejiang Province (119.676, 30.608, Lower Silurian)',
  'Reference (specimens firstly published)': 'Yang, D.Q. 1964. Some Lower Silurian graptolites from Anji, northwestern Zhejiang (Chekiang). Acta Palaeontologica Sinica, 12(4): 628-635.'
}
```

### Data Fields
- `Suborder` (string): Suborder of the graptoloidea.
- `Infraorder` (string): Infraorder of the graptoloidea.
- `Family (Subfamily)` (string): Family of the graptoloidea (with subfamily).
- `Genus` (string): genus of the graptoloidea
- `tagged species name` (string): The tagged name for the graptoloidea.
- `image` (string): File path for the image file.
- `Stage` (string): Which period is this graptoloidea in.
- `mean age value` (float): About how long has this specimen lasts.
- `Locality (Longitude, Latitude, Horizon)` (str): Where found the specimen (with longitude, latitude, and horizon).
- `Reference (specimens firstly published` (str): Who first publish this specimen.

### Data Splits
70% of data in training set; 15% of data in testing set; 15% of data in validation set.

## Dataset Creation

### Curation Rationale
The primary objective of curating the Graptoloidea Specimens dataset is to provide a comprehensive and accessible resource for the study and analysis of Graptoloidea, an order of extinct marine colonial organisms. This dataset is intended to support a wide range of scientific endeavors, including paleobiological research, evolutionary studies, and educational purposes in the field of paleontology. By assembling high-resolution images and detailed taxonomic information, the dataset aims to facilitate in-depth investigations into the morphology and classification of these ancient organisms, contributing to a broader understanding of Earth's geological history and biodiversity.

### Source Data
https://zenodo.org/records/6194943

### Data Processing
Specific processing explainition and code in CSV_Processing.py
## Bias, Risks, and Limitations
- **Technological and Methodological Constraints**: Utility is tied to current state of paleontological methodologies and technologies. Future advancements might necessitate reevaluation.
- **External Environmental Factors**: Might not fully account for the environmental conditions under which the specimens lived or were fossilized.
- **Imaging and Interpretation Limitations**: Quality and resolution of specimen images can vary, affecting morphological analyses. Interpretations based on images are subject to observer's expertise and bias.
## Citation
DOI: [10.5281/zenodo.5205215](https://doi.org/10.5281/zenodo.5205215)