GotThatData/kraken-trading-data

kraken-trading-data/
├── trades/                  # Individual trade executions
│   ├── XXBTZUSD_trades_2024-01.csv
│   ├── XETHZUSD_trades_2024-01.csv
│   └── ...
├── ohlcv/                   # OHLCV candlestick data (1min, 5min, 1h, 1d)
│   ├── XXBTZUSD_1min.csv
│   ├── XXBTZUSD_5min.csv
│   └── ...
├── orderbook/               # Limit order book snapshots
│   ├── XXBTZUSD_orderbook_2024-01.csv
│   └── ...
├── metadata.json            # Collection timestamps, pair info
└── README.md

from datasets import load_dataset

dataset = load_dataset("GotThatData/kraken-trading-data")

import pandas as pd

# Load Bitcoin trades for January 2024
df = pd.read_csv("hf://datasets/GotThatData/kraken-trading-data/trades/XXBTZUSD_trades_2024-01.csv")

# Convert timestamp to datetime
df['timestamp'] = pd.to_datetime(df['timestamp'], unit='ms')

# Set as index
df = df.set_index('timestamp')

# Display first rows
print(df.head())

                             price      volume  side         trade_id
timestamp                                                            
2024-01-01 00:00:01.234  42150.5    0.025000   buy   1704067201234-1
2024-01-01 00:00:01.567  42149.8    0.100000   sell  1704067201567-2
...

# Resample trades to hourly candles
ohlcv = df.resample('1H').agg({
    'price': ['first', 'max', 'min', 'last'],
    'volume': 'sum',
    'trade_id': 'count'
})

# Flatten column names
ohlcv.columns = ['open', 'high', 'low', 'close', 'volume', 'trade_count']
print(ohlcv.head())

import matplotlib.pyplot as plt

# Load 1-minute OHLCV
df = pd.read_csv("hf://datasets/GotThatData/kraken-trading-data/ohlcv/XXBTZUSD_1min.csv")
df['timestamp'] = pd.to_datetime(df['timestamp'], unit='ms')
df = df.set_index('timestamp')

# Calculate moving averages
df['MA_20'] = df['close'].rolling(window=20).mean()
df['MA_50'] = df['close'].rolling(window=50).mean()

# Plot
plt.figure(figsize=(14, 7))
plt.plot(df['close'], label='BTC Price', alpha=0.5)
plt.plot(df['MA_20'], label='20-period MA')
plt.plot(df['MA_50'], label='50-period MA')
plt.legend()
plt.title('Bitcoin Price with Moving Averages')
plt.show()

import seaborn as sns

# Load order book snapshot
ob = pd.read_csv("hf://datasets/GotThatData/kraken-trading-data/orderbook/XXBTZUSD_orderbook_2024-01.csv")

# Filter for a specific timestamp
snapshot = ob[ob['timestamp'] == ob['timestamp'].iloc[0]]

# Separate bids and asks
bids = snapshot[snapshot['side'] == 'bid'].sort_values('price', ascending=False)
asks = snapshot[snapshot['side'] == 'ask'].sort_values('price')

# Plot depth chart
plt.figure(figsize=(12, 6))
plt.step(bids['price'], bids['cumulative_volume'], where='pre', label='Bids', color='green')
plt.step(asks['price'], asks['cumulative_volume'], where='post', label='Asks', color='red')
plt.xlabel('Price (USD)')
plt.ylabel('Cumulative Volume (BTC)')
plt.title('BTC/USD Order Book Depth')
plt.legend()
plt.show()

from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split

# Load hourly OHLCV
df = pd.read_csv("hf://datasets/GotThatData/kraken-trading-data/ohlcv/XXBTZUSD_1h.csv")
df['timestamp'] = pd.to_datetime(df['timestamp'], unit='ms')
df = df.set_index('timestamp')

# Feature engineering
df['returns'] = df['close'].pct_change()
df['volatility'] = df['returns'].rolling(window=24).std()
df['volume_ma'] = df['volume'].rolling(window=24).mean()

# Target: 1 if price goes up next hour, 0 otherwise
df['target'] = (df['close'].shift(-1) > df['close']).astype(int)

# Drop NaNs
df = df.dropna()

# Features and target
features = ['open', 'high', 'low', 'close', 'volume', 'returns', 'volatility', 'volume_ma']
X = df[features]
y = df['target']

# Split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, shuffle=False)

# Train model
model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

# Evaluate
accuracy = model.score(X_test, y_test)
print(f"Accuracy: {accuracy:.2%}")

@dataset{daugherty2024kraken,
  author = {Bryan Daugherty},
  title = {Kraken Trading Data Collection},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/GotThatData/kraken-trading-data}
}