CarstenSchelp

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Ellipse
import matplotlib.transforms as transforms

def confidence_ellipse(x, y, ax, n_std=3.0, facecolor='none', **kwargs):
    """
    Create a plot of the covariance confidence ellipse of `x` and `y`
    
    See how and why this works: https://carstenschelp.github.io/2018/09/14/Plot_Confidence_Ellipse_001.html

CarstenSchelp

import numpy as np
from numpy import cos, sin
import matplotlib.pyplot as plt

"""
Looking at the covariance ellipse as a Lissajous figure.

In this gist: https://gist.github.com/CarstenSchelp/b992645537660bda692f218b562d0712
I plot a 2-D covariance ellipse by 
- normalizing the covariance matrix to contain only two ones, on the diagonal, and a pearson coefficient

CarstenSchelp

# The arctan2 funtion returns an angle between -PI and +PI (radians)
# If you need a range from zero to 2PI, this is how the
# arctan2 result can be converted.
# Zero still means "aligned with the 'x' axis"
# And the angle is still to be interpreted counter-clockwise.

import numpy as np

def arctan2_2pi(y, x):
    return -1 * (

CarstenSchelp

'''
A simple interactive file selection class based on IPython "Selection" widget.
Navigates through directories and displays all files that match the "file_pattern"
parameter.
'''
import ipywidgets as widgets
import glob
import os

class FileSelection:

CarstenSchelp

'''
Convert .net-style ticks to posix or python datetime.
.net ticks (like in C#) are 100 nanosecond counts
starting from january 1st in year 0001
This one is not very thorougly tested but worked for me when I needed it.
'''
import datetime

_epoch = datetime.date(1970, 1 , 1)
_dotnet_minvalue = datetime.date(1, 1, 1)

CarstenSchelp

import numpy as np

class OnlineCovariance:
    """
    A class to calculate the mean and the covariance matrix
    of the incrementally added data.
    The implementation is fully vectorized, which means that
    no for-loops or indices are involved when processing
    vectors and matrices.
    This gist attempts to cleanly implement and extend this

CarstenSchelp

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Ellipse
import matplotlib.transforms as transforms

def confidence_ellipse(x, y, ax, n_std=3.0, facecolor='none', **kwargs):
    """
    Create a plot of the covariance confidence ellipse of `x` and `y`
    
    See how and why this works: https://carstenschelp.github.io/2018/09/14/Plot_Confidence_Ellipse_001.html