A comprehensive, daily-updated Bitcoin dataset with OHLCV data and 70+ pre-calculated technical indicators. Perfect for machine learning, algorithmic trading, and financial analysis.
This dataset provides hourly Bitcoin price data enriched with a comprehensive suite of technical indicators calculated using TA-Lib. All indicators are pre-calculated and ready for immediate use in your models.
- 🕐 Hourly granularity - Detailed intraday data
- 🔄 Daily updates - Automated at midnight UTC
- 📊 70+ indicators - Moving averages, oscillators, volume, volatility, and more
- 🎯 ML-ready - Clean, structured CSV format
- 📉 Historical data - Complete Bitcoin trading history
- ⚡ Production-grade - Sourced from CryptoCompare API via Snowflake
- Filename:
bitcoin-hourly-technical-indicators.csv - Format: CSV (Comma-Separated Values)
- Encoding: UTF-8
- Size: ~XX MB (varies with historical data)
- Update Frequency: Daily at 00:05 UTC
| Column | Description | Type |
|---|---|---|
UNIX_TIMESTAMP |
Unix epoch timestamp | Integer |
DATETIME |
ISO 8601 datetime | Timestamp |
OPEN |
Opening price (USD) | Float |
HIGH |
Highest price (USD) | Float |
CLOSE |
Closing price (USD) | Float |
LOW |
Lowest price (USD) | Float |
VOLUME |
Trading volume (BTC) | Float |
- Simple Moving Averages (SMA): 5, 10, 20, 50, 100, 200 periods
- Exponential Moving Averages (EMA): 5, 10, 12, 20, 26, 50 periods
- Weighted Moving Averages (WMA): 10, 20 periods
- Double EMA (DEMA): 10, 20 periods
- Triple EMA (TEMA): 10, 20 periods
- Triangular MA (TRIMA): 20 periods
- Kaufman Adaptive MA (KAMA): 20 periods
- T3 Moving Average: 5 periods
- RSI (Relative Strength Index): 7, 14, 21 periods
- MACD (Moving Average Convergence Divergence): MACD, Signal, Histogram
- Stochastic Oscillator: SlowK, SlowD, FastK, FastD
- Stochastic RSI: FastK, FastD
- CCI (Commodity Channel Index): 14, 20 periods
- CMO (Chande Momentum Oscillator): 14 periods
- MOM (Momentum): 10 periods
- ROC (Rate of Change): 10 periods
- Williams %R: 14 periods
- PPO (Percentage Price Oscillator)
- APO (Absolute Price Oscillator)
- BOP (Balance of Power)
- Ultimate Oscillator (ULTOSC)
- AD (Accumulation/Distribution)
- ADOSC (Chaikin A/D Oscillator)
- OBV (On-Balance Volume)
- MFI (Money Flow Index): 14 periods
- Bollinger Bands: Upper, Middle, Lower
- ATR (Average True Range): 14 periods
- NATR (Normalized ATR): 14 periods
- TRANGE (True Range)
- TYPPRICE (Typical Price)
- ADX (Average Directional Index): 14 periods
- Directional Indicators: +DI, -DI, +DM, -DM
- Aroon Indicator: Up, Down, Oscillator
- SAR (Parabolic SAR)
- BETA (Beta coefficient)
- CORREL (Pearson Correlation)
- Linear Regression: Value, Angle, Slope
- STDDEV (Standard Deviation)
- HT_TRENDMODE (Trend vs Cycle Mode)
- HT_SINE (Sine Wave)
- HT_LEADSINE (Lead Sine Wave)
- HT_TRENDLINE (Instantaneous Trendline)
- MAMA (MESA Adaptive MA)
- Doji, Hammer, Inverted Hammer
- Hanging Man, Shooting Star
- Engulfing, Morning Star, Evening Star
- Three White Soldiers, Three Black Crows
- Harami, Dark Cloud Cover, Piercing
- Marubozu
- PRICE_CHANGE (Absolute price change)
- HIGH_LOW_RATIO (High/Low price ratio)
- CLOSE_OPEN_RATIO (Close/Open price ratio)
- VOLATILITY_30D (30-day rolling volatility)
- PRICE_VOLATILITY_30D (30-day price volatility)
# Clone the repository
git clone https://github.com/mouadja02/bitcoin-technical-indicators-dataset.git
cd bitcoin-technical-indicators-dataset
# Or download directly
wget https://raw.githubusercontent.com/mouadja02/bitcoin-technical-indicators-dataset/main/bitcoin-hourly-technical-indicators.csvimport pandas as pd
# Load the dataset
df = pd.read_csv('bitcoin-hourly-technical-indicators.csv')
# Convert datetime
df['DATETIME'] = pd.to_datetime(df['DATETIME'])
df.set_index('DATETIME', inplace=True)
# Display basic info
print(f"Dataset shape: {df.shape}")
print(f"Date range: {df.index.min()} to {df.index.max()}")
print(f"\nFirst few rows:\n{df.head()}")import matplotlib.pyplot as plt
# Plot Bitcoin price with moving averages
plt.figure(figsize=(14, 7))
plt.plot(df.index, df['CLOSE'], label='BTC Price', linewidth=1)
plt.plot(df.index, df['SMA_50'], label='SMA 50', linewidth=1.5)
plt.plot(df.index, df['SMA_200'], label='SMA 200', linewidth=1.5)
plt.title('Bitcoin Price with Moving Averages')
plt.xlabel('Date')
plt.ylabel('Price (USD)')
plt.legend()
plt.grid(True, alpha=0.3)
plt.show()
# RSI Analysis
plt.figure(figsize=(14, 5))
plt.plot(df.index, df['RSI_14'], label='RSI 14', color='purple')
plt.axhline(y=70, color='r', linestyle='--', label='Overbought')
plt.axhline(y=30, color='g', linestyle='--', label='Oversold')
plt.title('Relative Strength Index (RSI)')
plt.xlabel('Date')
plt.ylabel('RSI')
plt.legend()
plt.grid(True, alpha=0.3)
plt.show()from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
# Prepare features (select relevant indicators)
features = ['SMA_20', 'EMA_12', 'RSI_14', 'MACD', 'BB_UPPER',
'BB_LOWER', 'ATR_14', 'OBV', 'ADX_14', 'VOLUME']
X = df[features].dropna()
y = df['CLOSE'].shift(-1).loc[X.index] # Predict next hour's close
# Train-test split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.2, shuffle=False
)
# Train model
model = RandomForestRegressor(n_estimators=100, random_state=42)
model.fit(X_train, y_train)
# Evaluate
score = model.score(X_test, y_test)
print(f"R² Score: {score:.4f}")# Simple RSI strategy
df['signal'] = 0
df.loc[df['RSI_14'] < 30, 'signal'] = 1 # Buy signal
df.loc[df['RSI_14'] > 70, 'signal'] = -1 # Sell signal
# Calculate returns
df['returns'] = df['CLOSE'].pct_change()
df['strategy_returns'] = df['signal'].shift(1) * df['returns']
# Performance metrics
total_return = (1 + df['strategy_returns']).cumprod()[-1] - 1
print(f"Strategy Return: {total_return:.2%}")# Create custom features
df['price_momentum'] = df['CLOSE'] / df['SMA_20']
df['volatility_regime'] = pd.qcut(df['ATR_14'], q=3, labels=['Low', 'Medium', 'High'])
df['trend_strength'] = df['ADX_14'] > 25
# Volume analysis
df['volume_ma'] = df['VOLUME'].rolling(20).mean()
df['volume_surge'] = df['VOLUME'] > (df['volume_ma'] * 1.5)# Find bullish patterns
bullish_patterns = [
'CDL_HAMMER', 'CDL_INVERTED_HAMMER',
'CDL_MORNING_STAR', 'CDL_THREE_WHITE_SOLDIERS'
]
df['bullish_signal'] = df[bullish_patterns].sum(axis=1) > 0
# Find bearish patterns
bearish_patterns = [
'CDL_SHOOTING_STAR', 'CDL_HANGING_MAN',
'CDL_EVENING_STAR', 'CDL_THREE_BLACK_CROWS'
]
df['bearish_signal'] = df[bearish_patterns].sum(axis=1) > 0- ✅ No missing timestamps in sequence
- ✅ All fields are complete (No NULL values)
- Source: CryptoCompare API
- Calculations: TA-Lib (industry-standard library)
- Validation: Automated daily checks
CryptoCompare API → Apache Airflow → Snowflake DWH → GitHub
- Extraction: Hourly Bitcoin data from CryptoCompare
- Transformation: Technical indicators calculated using TA-Lib
- Loading: Stored in Snowflake data warehouse
- Publishing: Daily export to GitHub repository
- Frequency: Daily
- Time: 00:05 UTC
- Automation: Apache Airflow DAG
- Notification: Telegram bot alerts
- Uppercase: All column names are uppercase
- Underscore: Words separated by underscores
- Periods: Indicator periods specified with underscore (e.g.,
RSI_14)
- Books:
- "Technical Analysis of the Financial Markets" by John Murphy
- "Python for Finance" by Yves Hilpisch
- Courses:
- Coursera: Machine Learning for Trading
- Udemy: Algorithmic Trading with Python
Important: This dataset is for educational and research purposes only.
- ❌ NOT financial advice
- ❌ NOT investment recommendations
- ❌ NOT guaranteed to be error-free
Trading cryptocurrencies carries substantial risk. Past performance does not guarantee future results. Always do your own research and consult with financial professionals.
This dataset is released under the MIT License.
