pytalib

Pytalib is a python technical analysis library developed CMSC5720 project group which support various types of technical indicators. Pytalib adapts object oriented paradigm that each indicator is represented as an object. Unlike function-based library, using objects allow us to store some intermediate variables, for example Average gain/loss in RSI. This improves flexibility if we want to do further analysis on indicators.

Python version

Python 3.6.4

Dependencies

1. Networkx
2. Scipy

How to install

Pytalib has been published on Python Package Index (PyPi). Pytalib can be installed using the following command.

pip install pytalib

Types of indicators

Trend indicators

1. Moving Average Convergence Divergence
2. Simple Moving Average
3. Weighted Moving Average
4. Exponential Moving Average
5. Trix
6. Average Directional Index
7. Commodity Channel Index
8. Detrended Price Oscillator
9. Mass Index
10. Vortex Indicator

Momentum indicators

1. Rate of Change
2. Relative Strength Index
3. Stochastic Oscillator
4. Money Flow Index
5. True Strength Index
6. Ultimate Oscillator
7. Williams
8. Know Sure Thing Oscillator

Volatility indicators

1. Average True Range
2. Bollinger Bands
3. Price Channel
4. Keltner Channel
5. Standard Deviation

Volume indicators

1. Accumulation Distribution Line
2. Ease of Movement
3. Force Index
4. Negative Volume Index
5. On Balance Volume
6. Put Call Ratio

Visibility Graph Algorithm

Implementations the following time series-to-graph algorithm which takes the time series as parameter and returns a networkx undirected graph.

1. ts2vg_basic(series)

Reference: "From time series to complex networks: The visibility graph" by L. Lacasa, B. Luque, F. Ballesteros, J. Luque, and J. C. Nuno

2. ts2vg_fast(series)

Reference: "Fast transformation from time series to visibility graphs" by Xin Lan, Hongming Mo, Shiyu Chen, Qi Liu, and Yong Deng

3. ts2hvg(series)

Reference: "Horizontal visibility graphs: exact results for random time series" by B. Luque , L. Lacasa, F. Ballesteros and J. Luque

Correlation analysis

Implementation of multiscale horizontal-visibility-graph correlation analysis (MHVGCA) that utilised horizontal visibility graph and degree sequence similarity to estimate the correlation between time series under specific time scale.

1. mhvgca_method(series_a, series_b, timescale=20)

Reference: "Multiscale horizontal-visibility-graph correlation analysis of stock time series" by Weidong Li and Xiaojun Zhao

Example Code

Calculate indicators

from pytalib.indicators.trend import SimpleMovingAverage

prices = [1,2,3,4,5,6,7,8,9,10]
sma = SimpleMovingAverage(prices=prices, period=3)
result = sma.calculate()

# reuse sma object
prices2 = [10,9,8,7,6,5,4,3,2,1]
sma.reset(prices=prices2, period=3)
result2 = sma.calculate()

Time series-to-Graph transformation

import networkx as nx
from pytalib.graph import visibility_graph as vg
import matplotlib.pyplot as plt
import matplotlib
matplotlib.use('TkAgg')

prices = [1,3,2,4,5,6,9,8,9,10]
G = vg.ts2vg_fast(prices)
nx.draw_networkx(G, with_labels=True, font_weight='bold')
plt.title('visibility graph of prices')
plt.show()

import networkx as nx
from pytalib.graph import visibility_graph as vg
import matplotlib.pyplot as plt
import matplotlib
matplotlib.use('TkAgg')

prices = [1,3,2,4,5,6,9,8,9,10]
G = vg.ts2hvg(prices)
nx.draw_networkx(G, with_labels=True, font_weight='bold')
plt.title('horizontal visibility graph of prices')
plt.show()