Sign in Get started

DATA SCIENCE MACHINE LEARNING PROGRAMMING VISUALIZATION AI JOURNALISM PICKS CONTRIBUTE

Towards Data
Science
Sharing concepts,
Implementing MACD in Python
ideas, and codes. MACD is a popularly used technical indicator in trading stocks,
Follow currencies, cryptocurrencies, etc.
Luke Posey Follow
Mar 30 · 5 min read
255

MACD is popularly used in analyzing charts for stocks, currencies, crypto, and other assets…Credit: Unsplash

Basics of MACD
MACD is used and discussed in many different trading circles. Moving Average
Convergence Divergence (MACD) is a trend following indicator. MACD can be
calculated very simply by subtracting the 26 period EMA from the 12 period
EMA. We previously discussed EMAs in our article here. MACD can be used
and interpreted in a handful of different ways to give the trader potential
value and insight into their trading decisions.

Useful Strategies

MACD is commonly used by analyzing crossovers, divergences, and periods of

steep slope (positive or negative). Along with the MACD line (from
subtracting the 12 period EMA from the 16 period EMA) the chart commonly
will include a signal line plotted on top of the MACD. This signal line is a 9 day
EMA of the MACD.

In a bullish crossover, just like in Moving Averages, a buy signal occurs when
MACD crosses above the signal line. A bearish signal occurs when MACD
crosses below the signal line. If a crossover occurs with a high sloping MACD,
this can be a sign of an overbought or oversold condition, depending on if the
crossover is bullish or bearish respectively. MACD is a great indicator for
understanding if movement in the price is strong or weak. A weak movement
is likely to correct and a strong movement is likely to continue.

Divergences are also simple to understand. When the MACD establishes a

high or low diverging from highs or lows in the price it establishes a
divergence. A bullish divergence is in place when MACD has two rising lows
on the MACD with two falling lows on the asset price. Divergences can be
used to find a changing trend. Traders are always looking for the competitive
edge and predicting a trend change can be very profitable. Of course,
divergences are not completely reliable and should only be used as an
additional piece of information, not a sole indication of price direction.

Steep slope can signal an overbought or oversold situation. In such a situation

a stock’s trend is likely soon to lose steam and see a correction or reversal from
current direction.

Python Implementation
We start as we always do by picking a stock and gathering the data. As usual,
we’ll do our analysis on AMD and use the IEX API to grab the data. IEX has
always been a highly reliable source of data for me, but you can use Quandl or
whatever other source of data you prefer.

import pandas as pd
import numpy as np
from datetime import datetime
import matplotlib.pyplot as plt
import pyEX as p

ticker = 'AMD'
timeframe = '6m'

df = p.chartDF(ticker, timeframe)
df = df[['close']]
df.reset_index(level=0, inplace=True)
df.columns=['ds','y']

plt.plot(df.ds, df.y, label='AMD')

plt.show()

AMD from late 2018 to present date (early 2019).

exp1 = df.y.ewm(span=12, adjust=False).mean()

exp2 = df.y.ewm(span=26, adjust=False).mean()

macd = exp1-exp2
exp3 = macd.ewm(span=9, adjust=False).mean()

plt.plot(df.ds, macd, label='AMD MACD', color = '#EBD2BE')

plt.plot(df.ds, exp3, label='Signal Line', color='#E5A4CB')
plt.legend(loc='upper left')
plt.show()

This allows us to plot the MACD vs the signal line. See if you can spot the
bullish and bearish crossovers!

MACD vs Signal Line

Check the graph below. Were you correct? Remember, a bullish crossover
happens when the MACD crosses above the signal line and a bearish crossover
happens when the MACD crosses below the signal line.

Bullish crossover represented in green, bearish crossover represented in red.

The above example was a simple way to use MACD to study crossovers. Next,
let’s study strength and examine overbought or oversold conditions.

We start by implementing the exponential moving averages and MACD.

exp1 = df.y.ewm(span=12, adjust=False).mean()

exp2 = df.y.ewm(span=26, adjust=False).mean()
exp3 = df.y.ewm(span=9, adjust=False).mean()

macd = exp1-exp2

plt.plot(df.ds, df.y, label='AMD')

plt.plot(df.ds, macd, label='AMD MACD', color='orange')
plt.plot(df.ds, exp3, label='Signal Line', color='Magenta')
plt.legend(loc='upper left')
plt.show()

Blue line represents the AMD stock price, orange line represents the MACD

We can blow this MACD line up a bit by plotting it separate from the stock
price and see the steep slopes more clearly.

MACD from late 2018 to present date (early 2019).

Let’s recall our discussion of overbought and oversold from earlier. We can see
the MACD stays pretty flat over time. But there are certain times where the
MACD curve is steeper than others. These are instances of overbought or
oversold conditions. We represent our oversold conditions with green circles
and overbought with red circles. You can see that soon after the MACD shows
an overbought or oversold condition the momentum slowed and the stock
price reacted accordingly.

Green circles correspond to bullish divergence, red corresponds to bearish divergence

We briefly discussed MACD and implemented it in Python to examine its use

in crossovers and overbought/oversold conditions. Hopefully this article
helped you add another tool to your trading toolbox!

For a more detailed dive into MACD:

MACD - Wikipedia
The average series is also a derivative estimate, with an additional low-pass
lter in tandem for further smoothing…

en.wikipedia.org

Moving Average Convergence Divergence (MACD)

Moving Average Convergence Divergence (MACD) is a trend-following
momentum indicator that shows the relationship…

www.investopedia.com

Trading Data Science Cryptocurrency Programming Python

255 claps

Luke Posey Follow Towards Data Follow

Medium member since Science
Feb 2019 Sharing concepts, ideas,
I am a developer, student, and codes.
and engineer. I like
statistics and enjoy
applying AI/ML to
healthcare, markets,
controls, etc. 👨🏻💻

More from Towards Data Science More from Towards Data Science More from Towards Data Science
10 Simple hacks to speed up your Why you’re not a job-ready data What 70% of Data Science Learners
Data Analysis in Python scientist (yet) Do Wrong

Parul Pandey Jeremie Harris Dan Becker

5.2K 7K 3.2K
Jun 17 · 8 min read Jun 17 · 6 min read Jun 14 · 3 min read

Responses

Write a response…