Pandas Extension¶

numta provides a .ta accessor for Pandas DataFrames, making it easy to calculate technical indicators directly on your data.

Setup¶

The accessor is automatically registered when you import numta:

import pandas as pd
import numta  # Auto-registers the .ta accessor

Basic Usage¶

Creating a DataFrame¶

import pandas as pd
import numpy as np
import numta

# Create sample OHLCV data
np.random.seed(42)
n = 100
df = pd.DataFrame({
    'open': np.random.uniform(100, 110, n),
    'high': np.random.uniform(110, 120, n),
    'low': np.random.uniform(90, 100, n),
    'close': np.random.uniform(100, 110, n),
    'volume': np.random.randint(1000, 10000, n)
})

Calculating Indicators¶

Return as Series¶

# Calculate indicators and return as Series
sma = df.ta.sma(timeperiod=20)
ema = df.ta.ema(timeperiod=12)
rsi = df.ta.rsi(timeperiod=14)

Append to DataFrame¶

# Calculate and append to DataFrame
df.ta.sma(timeperiod=20, append=True)   # Adds 'SMA_20' column
df.ta.ema(timeperiod=12, append=True)   # Adds 'EMA_12' column
df.ta.rsi(timeperiod=14, append=True)   # Adds 'RSI_14' column

Available Methods¶

Moving Averages¶

Method	Description
`df.ta.sma(timeperiod)`	Simple Moving Average
`df.ta.ema(timeperiod)`	Exponential Moving Average
`df.ta.wma(timeperiod)`	Weighted Moving Average
`df.ta.dema(timeperiod)`	Double Exponential MA
`df.ta.tema(timeperiod)`	Triple Exponential MA
`df.ta.kama(timeperiod)`	Kaufman Adaptive MA
`df.ta.trima(timeperiod)`	Triangular MA
`df.ta.t3(timeperiod)`	T3 Moving Average

Momentum Indicators¶

Method	Description
`df.ta.rsi(timeperiod)`	Relative Strength Index
`df.ta.macd(fastperiod, slowperiod, signalperiod)`	MACD
`df.ta.stoch(fastk_period, slowk_period, slowd_period)`	Stochastic
`df.ta.adx(timeperiod)`	Average Directional Index
`df.ta.cci(timeperiod)`	Commodity Channel Index
`df.ta.mfi(timeperiod)`	Money Flow Index
`df.ta.mom(timeperiod)`	Momentum
`df.ta.roc(timeperiod)`	Rate of Change
`df.ta.willr(timeperiod)`	Williams %R

Volatility Indicators¶

Method	Description
`df.ta.atr(timeperiod)`	Average True Range
`df.ta.natr(timeperiod)`	Normalized ATR
`df.ta.trange()`	True Range

Bands¶

Method	Description
`df.ta.bbands(timeperiod, nbdevup, nbdevdn)`	Bollinger Bands

Volume Indicators¶

Method	Description
`df.ta.obv()`	On Balance Volume
`df.ta.ad()`	Accumulation/Distribution
`df.ta.adosc(fastperiod, slowperiod)`	A/D Oscillator

Trend Indicators¶

Method	Description
`df.ta.sar(acceleration, maximum)`	Parabolic SAR

Pattern Recognition¶

Method	Description
`df.ta.cdldoji()`	Doji Pattern
`df.ta.cdlengulfing()`	Engulfing Pattern
`df.ta.cdlhammer()`	Hammer Pattern
`df.ta.find_patterns()`	Find all patterns
`df.ta.find_harmonic_patterns()`	Find harmonic patterns

Multi-Output Indicators¶

Some indicators return multiple outputs:

# MACD returns three columns
df.ta.macd(append=True)
# Adds: MACD_12_26_9, MACDsignal_12_26_9, MACDhist_12_26_9

# Bollinger Bands returns three columns
df.ta.bbands(append=True)
# Adds: BBU_5_2, BBM_5_2, BBL_5_2

# Stochastic returns two columns
df.ta.stoch(append=True)
# Adds: STOCHk_5_3_0_3_0, STOCHd_5_3_0_3_0

Column Naming Convention¶

When using append=True, columns are named with the format:

{INDICATOR}_{PARAM1}_{PARAM2}...

Examples: - SMA_20 - SMA with timeperiod=20 - EMA_12 - EMA with timeperiod=12 - RSI_14 - RSI with timeperiod=14 - MACD_12_26_9 - MACD with fast=12, slow=26, signal=9

Custom Column Names¶

You can specify custom column names with the column parameter:

df.ta.sma(timeperiod=20, append=True, column='my_sma')
# Adds 'my_sma' column instead of 'SMA_20'

Working with Different Column Names¶

If your DataFrame uses different column names:

df = pd.DataFrame({
    'price': prices,
    'hi': highs,
    'lo': lows,
    'vol': volumes
})

# Specify column mappings
df.ta.set_columns(close='price', high='hi', low='lo', volume='vol')

# Now indicators use the correct columns
df.ta.sma(timeperiod=20)

Complete Example¶

import pandas as pd
import numpy as np
import numta

# Create sample data
np.random.seed(42)
n = 200
df = pd.DataFrame({
    'open': np.random.uniform(100, 110, n),
    'high': np.random.uniform(110, 120, n),
    'low': np.random.uniform(90, 100, n),
    'close': np.random.uniform(100, 110, n),
    'volume': np.random.randint(1000, 10000, n)
})

# Add multiple indicators
df.ta.sma(timeperiod=20, append=True)
df.ta.sma(timeperiod=50, append=True)
df.ta.ema(timeperiod=12, append=True)
df.ta.rsi(timeperiod=14, append=True)
df.ta.macd(append=True)
df.ta.bbands(append=True)
df.ta.atr(timeperiod=14, append=True)

# View results
print(df.tail())

# Generate signals
df['signal'] = 0
df.loc[(df['SMA_20'] > df['SMA_50']) & (df['RSI_14'] < 30), 'signal'] = 1
df.loc[(df['SMA_20'] < df['SMA_50']) & (df['RSI_14'] > 70), 'signal'] = -1

# Filter signals
signals = df[df['signal'] != 0]
print(f"Number of signals: {len(signals)}")