06-summing-up.md

layout	title	subtitle	minutes
page	Advanced NumPy	Summing up	30

Quantile normalisation {.challenge}

Quantile Normalization is a method to align distributions. Implement it using NumPy axis-wise operations and fancy indexing.

Hint: look for documentation for np.sort, and np.argsort.*

def qnorm(x):
    """Quantile normalize an input matrix.
    
    Parameters
    ----------
    x : 2D array of float, shape (M, N)
        The input data, with each column being a
        distribution to normalize.
        
    Returns
    -------
    xn : 2D array of float, shape (M, N)
        The normalized data.
    """
    return x 

import numpy as np
data = np.array([[5, 4, 3],
                 [2, 1, 4],
                 [3, 5, 6],
                 [4, 2, 7]])
from numpy.testing import assert_array_almost_equal
assert_array_almost_equal(qnorm(data), results, decimal=2)