Overview

Use PeakUtils to find peaks in a graph. For more informations about different peaks implementations, see https://blog.ytotech.com/2015/11/01/findpeaks-in-python.

Update 2019-04-11: A better way to find peaks is to use scipy.signal.argrelextrema() function.

Requirement

apple.csv containing Apple stock prices data. See sample data section below for more details. Each column is separated by a tab.
You need mpl_finance Python package to draw candlestick chart. See here for instructions on how to install mpl_finance.

Installation

# Download PeakUtils zip file from https://bitbucket.org/lucashnegri/peakutils/downloads/
pip install lucashnegri-peakutils-XXXXXX.zip

Usage

Look at the code below at the Peak section. It is very simple to use the PeakUtils library.

#!/usr/bin/python3
 
import matplotlib
matplotlib.use('Agg') # Bypass the need to install Tkinter GUI framework
 
# Avoid FutureWarning: Pandas will require you to explicitly register matplotlib converters.
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()
 
import pandas as pd
from mpl_finance import candlestick_ohlc
 
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
 
import peakutils
 
# Load data from CSV file.
##########################
my_headers = ['date', 'open', 'high', 'low', 'close', 'volume']
my_dtypes = {'date': 'str', 'open': 'float', 'high': 'float', 'low': 'float', 
                                                'close': 'float', 'volume': 'int'}
my_parse_dates = ['date']
loaded_data = pd.read_csv('apple.csv', sep='\t', header=1, names=my_headers, 
                            dtype=my_dtypes, parse_dates=my_parse_dates)
 
 
# Convert 'Timestamp' to 'float'.
#   candlestick_ohlc needs time be in float days format - see date2num.
loaded_data['date'] = [mdates.date2num(d) for d in loaded_data['date']]
 
# Re-arrange so that each line contains values of a day: 'date','open','high','low','close'.
quotes = [tuple(x) for x in loaded_data[['date','open','high','low','close']].values]
 
# Plot candlestick.
##########################
fig, ax = plt.subplots()
candlestick_ohlc(ax, quotes, width=0.5, colorup='g', colordown='r');
 
# Peak
############
# Find peak points from high values.
peak_indexes = peakutils.indexes(loaded_data['high'].values, thres=0.5, min_dist=30)
peak_x = loaded_data['date'].values[peak_indexes]
peak_y = loaded_data['high'].values[peak_indexes]
 
plt.plot(peak_x, peak_y, "x") # Mark 'X' as peaks.
 
 
# Customize graph.
##########################
plt.xlabel('Date')
plt.ylabel('Price')
plt.title('Apple - Peaks are marked as X')
 
# Format time.
ax.xaxis_date()
ax.xaxis.set_major_formatter(mdates.DateFormatter("%Y-%m-%d"))
 
plt.gcf().autofmt_xdate()   # Beautify the x-labels
plt.autoscale(tight=True)
 
# Save graph to file.
plt.savefig('peak.png')

Chart output

Chart show peaks

apple.csv sample data

date open    high    low close   volume
2001-01-02  1.062500    1.089286    1.040179    1.062500    113078000
2001-01-03  1.035714    1.191964    1.031250    1.169643    204268400
2001-01-04  1.295757    1.321429    1.200893    1.218750    184849000
2001-01-05  1.209821    1.241071    1.147321    1.169643    103089000
2001-01-08  1.209821    1.213164    1.138393    1.183036    93424800

Python

About the author

Xuan Ngo is the founder of OpenWritings.net. He currently lives in Montreal, Canada. He loves to write about programming and open source subjects.

Python - Find peaks in a graph using PeakUtils