ZigZag provides functions for identifying the peaks and valleys of a time series. Additionally, it provides a function for computing the maximum drawdown.

If numba can be imported, it will be used to greatly accelerate the execution time at the cost of a small initial compile. On my machine, the JIT'd code executed about 38 times faster than the pure Python code. Numba is not an install_requires package because numba can be a pain to install for some people. Learn more about numba here.

For fastest understanding, view the IPython notebook demo tutorial.

This is an admittedly small project. Still, if you have any contributions, please fork this project on github and send me a pull request. You can email me, John B Nelson, at jbn@pathdependent.com with any questions.