[TA-Lib] #6: Analyzing and Calculating MACD with TA-Lib

MACD (Moving Average Convergence Divergence) is one of the methods used to analyze the past data of a market to predict future price movements. It's widely utilized for analyzing market trends and momentum. In this article, we will introduce the process of calculating the MACD Line, Signal Line, and Histogram using "TA-Lib". We will also thoroughly examine how to use MACD for trend analysis, and the criteria for determining overbought/oversold conditions. You can find how to install TA-Lib in [TA-Lib] #2: Installing TA-Lib for Technical Analysis, and the basic concept of MACD in [TA-Lib] #5 MACD - Analyzing MACD Meaning and Its Role in Market Trends.

Calculating MACD Line, Signal Line, and Histogram Using TA-Lib

The calculation of MACD (Moving Average Convergence Divergence) can be effortlessly performed using the TA-Lib library.

Calculating MACD with Basic Settings

The default formula for MACD in TA-Lib involves the difference between the long-term (typically 26 days) exponential moving average (EMA) and the short-term (typically 12 days) EMA. The Signal Line is composed of the 9-day EMA of this MACD Line. The Histogram, being the difference between the MACD Line and Signal Line, is used to evaluate the market's momentum.

python
# Importing the library
from yahooquery import Ticker
import talib
import pandas as pd

# Fetching Google (GOOG) stock data
google = Ticker('GOOG')
df = google.history(period='2y')  # 2-year data

# Calculating MACD with basic settings
macd_line, signal_line, histogram = talib.MACD(df['close'])

# Adding results to DataFrame
df['macd'] = macd_line
df['signal'] = signal_line
df['histogram'] = histogram

# Printing the results
print(df.tail())
                macd    signal  histogram
date                                     
2023-08-17  1.801564  2.071457  -0.269893
2023-08-18  1.525293  1.962224  -0.436931
2023-08-21  1.356871  1.841153  -0.484282
2023-08-22  1.270081  1.726939  -0.456858
2023-08-23  1.468408  1.675233  -0.206825

Calculating MACD with Custom Settings

If you wish to set the periods for the moving averages customly, you can adjust the fastperiod, slowperiod, signalperiod parameters in TA-Lib's MACD function.

python
# Calculating MACD with custom moving average periods
# For example, changing fast_period to 10, and slow_period to 22:
macd_line_custom, signal_line_custom, histogram_custom = talib.MACD(df['close'], fastperiod=10, slowperiod=22)

# Adding results to DataFrame
df['macdCustom'] = macd_line_custom
df['signalCustom'] = signal_line_custom
df['histogramCustom'] = histogram_custom

# Printing the results
print(df.tail())
            macdCustom  signalCustom  histogramCustom
date                                                 
2023-08-17    1.472135      1.842813        -0.370678
2023-08-18    1.156198      1.705490        -0.549292
2023-08-21    0.979688      1.560330        -0.580642
2023-08-22    0.904435      1.429151        -0.524716
2023-08-23    1.167833      1.376887        -0.209054

Through the above code, you can perform MACD analysis by adjusting the periods of each moving average to suit various market conditions and strategies.

MACD Analysis Using TA-Lib

Trend Following

Identifying MACD Line and Signal Line Crossover Points

The crossover between the MACD Line and Signal Line serves as an important signal for recognizing market fluctuations. By capturing these crossovers, you can discern the possibility of a trend change and make investment decisions accordingly.

  • Golden Cross: This occurs when the MACD Line crosses the Signal Line from below. This phenomenon often represents the commencement of an uptrend, and can be interpreted as an opportunity to buy.
  • Dead Cross: This happens when the MACD Line crosses the Signal Line from above. This signal may indicate the beginning of a downtrend, and can be considered a point to sell.
python
import talib

macd, signal, _ = talib.MACD(df['close'])
df['macd'] = macd
df['signal'] = signal

df['upward'] = df['macd'] > df['signal']
df['downward'] = df['macd'] < df['signal']
df['goldenCross'] = (df['upward']) & (~df['upward'].shift(1).fillna(False))
df['deadCross'] = (df['downward']) & (~df['downward'].shift(1).fillna(False))

print('Golden Cross points:')
print(df[df['goldenCross']])

print('Dead Cross points:')
print(df[df['deadCross']])
Golden Cross points:
            goldenCross  deadCross
date                              
2021-10-11         True      False
...
2023-07-26         True      False


Dead Cross points:
            goldenCross  deadCross
date                              
2021-11-22        False       True
...
2023-08-15        False       True

Analyzing Trends Using the Size of the Histogram

Trend analysis using the size of the Histogram can be conducted in two significant ways:

  • Strong Trend: If the size of the Histogram is progressively increasing, it signifies that the current trend (either upward or downward) is intensifying. This represents a strong force in the market and can enhance investment strategies in that direction.
  • Trend Weakening: If the size is diminishing, the trend is weakening, indicating that it might be time to adjust positions. This implies a need to be wary of trend reversals and respond appropriately.
python
macd, signal, histogram = talib.MACD(df['close'])
df['macd'] = macd
df['signal'] = signal
df['histogram'] = histogram

# Initializing trend columns
df['strong_bull'] = False
df['strong_bear'] = False
df['weak_bull'] = False
df['weak_bear'] = False

consecutive_increase = 0
consecutive_decrease = 0
prev_hist_size = abs(df['histogram'][0])

for i in df.index:
    curr_hist_size = abs(df['histogram'][i])

    if curr_hist_size > prev_hist_size:
        consecutive_increase += 1
        consecutive_decrease = 0
    elif curr_hist_size < prev_hist_size:
        consecutive_decrease += 1
        consecutive_increase = 0

    if consecutive_increase >= 5:
        trend_col = 'strong_bull' if df['histogram'][i] > 0 else 'strong_bear'
        df.at[i, trend_col] = True
        consecutive_increase = 0

    if consecutive_decrease >= 5:
        trend_col = 'weak_bull' if df['histogram'][i] > 0 else 'weak_bear'
        df.at[i, trend_col] = True
        consecutive_decrease = 0

    prev_hist_size = curr_hist_size

print('Starting points of strong bullish trends:')
print(df[df['strong_bull']])

print('Starting points of strong bearish trends:')
print(df[df['strong_bear']])

print('Starting points of weakening bullish trends:')
print(df[df['strong_bear']])

print('Starting points of weakening bearish trends:')
print(df[df['weak_bear']])
Starting points of strong bullish trends:
            strong_bull  strong_bear  weak_bull  weak_bear
date                                                      
2021-10-19         True        False      False      False
...
2023-05-16         True        False      False      False


Starting points of strong bearish trends:
            strong_bull  strong_bear  weak_bull  weak_bear
date                                                      
2022-01-07        False         True      False      False
...
2023-08-18        False         True      False      False


Starting points of weakening bullish trends:
            strong_bull  strong_bear  weak_bull  weak_bear
date                                                      
2021-10-26        False        False       True      False
...
2023-05-30        False        False       True      False


Starting points of weakening bearish trends:
            strong_bull  strong_bear  weak_bull  weak_bear
date                                                      
2021-12-08        False        False      False       True
...
2023-07-05        False        False      False       True

This code analyzes the trends by tracking five or more consecutive increases or decreases in the size of the Histogram. By following this method, market trends can be more accurately discerned, enabling precise strategy adjustments.

Overbought/Oversold

MACD can be used to determine the overbought and oversold conditions of a market. By analyzing the divergence between the MACD Line and the Signal Line or calculating the crossing with the zero line, one can discern the market state.

Divergence Calculation Between MACD Line and Signal Line

The divergence between the MACD Line and Signal Line can be seen as the distance between the two lines, allowing us to analyze the market condition. A wide distance may indicate that the market is overheated (overbought), potentially signaling a selling opportunity. Conversely, if the distance is narrow, it could signify an oversold market condition, potentially providing a buying opportunity.

The criteria for divergence may vary depending on market conditions and the characteristics of the asset. In the code example below, if the current divergence is 80% or more of the maximum divergence in the past 30 days, it is considered that the two lines are too far apart and is deemed a selling timing. Conversely, if the current divergence is less than 120% of the minimum divergence in the past 30 days, it's considered that the two lines are too close, and is deemed a buying timing.

python
macd, signal, _ = talib.MACD(df['close'])
df['macd'] = macd
df['signal'] = signal

# Calculate the divergence between MACD Line and Signal Line.
df['divergence'] = df['macd'] - df['signal']

# Calculate the maximum and minimum divergence in the past 30 days.
df['max_divergence'] = df['divergence'].rolling(window=30).max()
df['min_divergence'] = df['divergence'].rolling(window=30).min()

# Sell timing determination: Current divergence is 80% or more of the maximum divergence.
df['sell_signal'] = df['divergence'] >= df['max_divergence'] * 0.8

# Buy timing determination: Current divergence is less than 120% of the minimum divergence.
df['buy_signal'] = df['divergence'] <= df['min_divergence'] * 1.2

print('Overbought')
print(df[df['sell_signal']])

print('Oversold')
print(df[df['buy_signal']])
Overbought
            sell_signal  buy_signal
date                               
2022-02-01         True       False
...                
2023-08-01         True       False


Oversold
            sell_signal  buy_signal
date                               
2021-11-19        False        True

80% and 120% are experimental values and can be adjusted according to individual market conditions and strategies.

Zero Line Crossing Calculation

When the MACD Line crosses the zero line upward or downward, it can be an important trading signal. An upward crossing may signal a buying opportunity, while a downward crossing may signal a selling opportunity.

python
# Calculate zero line crossing
df['cross_zero_line_up'] = (df['macd'] > 0) & (df['macd'].shift(1) <= 0)
df['cross_zero_line_down'] = (df['macd'] < 0) & (df['macd'].shift(1) >= 0)

print('Crossing Zero Line Upward')
print(df[df['cross_zero_line_up']])

print('Crossing Zero Line Downward')
print(df[df['cross_zero_line_down']])
Crossing Zero Line Upward
            cross_zero_line_up  cross_zero_line_down
date                                                
2021-10-20                True                 False
...
2023-07-14                True                 False


Crossing Zero Line Downward
            cross_zero_line_up  cross_zero_line_down
date                                                
2021-10-22               False                  True
...
2023-07-10               False                  True

MACD is an excellent tool for trend and momentum analysis, serving as an essential indicator in investment decisions. Through this article, you have learned how to effortlessly calculate and analyze MACD using TA-Lib. This enables you to conduct analysis tailored to various market situations and strategies, leading to more precise investment decisions.

FAQs

  1. Can I adjust parameters other than fastperiod, slowperiod, and signalperiod in TA-Lib's MACD function?
    • The MACD function in TA-Lib typically allows adjustment of these three key parameters. Additional settings are not generally used unless in special cases.
  2. How should I choose the periods used in MACD analysis?
    • The periods used in MACD analysis can vary depending on the strategy and market situation. Common default values are 12, 26, and 9 days, but the optimal values for specific markets or products can be found through experimentation.
  3. How should I interpret a negative MACD Histogram?
    • If the MACD Histogram is negative, it indicates that the MACD Line is lower than the Signal Line. This suggests a higher likelihood of a downtrend, and it may be interpreted as a selling signal. Careful analysis is required as interpretation may vary according to specific market conditions and strategies.
  4. How can I perform MACD analysis using yahooquery?
  5. What should I do if I want to analyze MACD without installing TA-Lib?
    • It is possible to calculate MACD without TA-Lib, although the code may become more complex. You can implement it directly using libraries like Pandas in Python. However, it is recommended to install TA-Lib to enjoy its convenience and efficiency. Information on installation can be found in [TA-Lib] #2: Installing TA-Lib for Technical Analysis.
© Copyright 2023 CLONE CODING