sogosonnet/Crypto-Chart-Dataset

---
license: cc-by-4.0
task_categories:
- image-classification
language:
- en
tags:
- finance
- cryptocurrency
- candlestick-chart
- technical-analysis
size_categories:
- 10K<n<100K
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': down_7plus
'1': down_5_7
'2': down_3_5
'3': up_3_5
'4': up_5_7
'5': up_7plus
- name: ticker
dtype: string
- name: end_date
dtype: string
- name: pct_return
dtype: float32
splits:
- name: train
num_bytes: 362633299
num_examples: 18549
- name: test
num_bytes: 207396726
num_examples: 10474
download_size: 572139048
dataset_size: 570030025
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
---

# Crypto-Chart-Dataset

Candlestick chart image dataset for cryptocurrency price prediction research.

## Overview

| Item | Value |
|------|-------|
| **Tickers** | 15 (BTC, ETH, BNB, XRP, ADA, DOGE, LTC, LINK, XLM, TRX, ETC, XMR, NEO, EOS, IOTA) |
| **Total Images** | 29,080 |
| **Period** | 2015 – 2025 |
| **Image Size** | ~480×480 px (4×4 inch, 120 dpi) |
| **Labels** | 6-class forward return (±3%/±5%/±7%, adjusted for crypto volatility) |

## Download

Download `crypto_images.zip` from the Files tab and extract:

```bash
unzip crypto_images.zip
```

## Chart Specification

Same as [SP500-Chart-Dataset](https://huggingface.co/datasets/sogosonnet/SP500-Chart-Dataset):

- OHLCV candlestick (Charles style) + Volume bars
- MA5 (blue dotted), MA60 (red dashed), MA120 (green solid)
- Bollinger Bands (grey shaded area, α=0.15)
- Percentage-normalized prices (scale-invariant across tickers)

## Label Definition

Thresholds adjusted for higher crypto volatility:

| Label | Return Range |
|-------|-------------|
| `down_7plus` | r < −7% |
| `down_5_7` | −7% ≤ r < −5% |
| `down_3_5` | −5% ≤ r < −3% |
| `up_3_5` | 3% < r ≤ 5% |
| `up_5_7` | 5% < r ≤ 7% |
| `up_7plus` | r > 7% |

> Returns in [−3%, +3%] are excluded as ambiguous.

## Citation

```bibtex
@misc{sp500chart2025,
title={SP500-Chart-Dataset},
author={Ahn, Jaehyun},
year={2025},
howpublished={\url{https://github.com/JaehyunAhn/SP500-Chart-Dataset}},
note={Yonsei University}
}
```

## License

CC BY 4.0