npzigzag 0.3

Vanilla numpy implementation of zigzag indicator.

npzigzag

This is a simple numpy implementation of a zigzag indicator. It's only dependencies are pandas and numpy.

Usage

pip install npzigzag

from npzigzag import core as zz

This package has one method:

core.zigzag(X, pc, include_first = True)

Parameters:

• X: pd.Series or np.ndarray or list
  • Main data to compute the indicator on.
• pc: float
  • Percentage change treshold. This indicates the minimum jump the timeseries has to make before a pivot is recognized in the zigzag indicator.
• include_first: bool
  • Indicates wether or not to add the first observation of X as a pivot point,

Returns:

• pivot_indices: np.ndarray if X is np.ndarray or list, pd.Int64Index if X is pd.Series
  • Array with indices of zigzag pivot points.

Example:

from npzigzag import core as zz
import numpy as np

X = np.cumprod(1 + np.random.randn(100)/100)
zz_pivots = zz.zigzag(X,0.02, False)

plt.plot(X, '--')
plt.plot(zz_pivots, X[zz_pivots])
plt.scatter(zz_pivots, X[zz_pivots])

Ouputs: