CODEWITHSUNDEEP
pythondata-science
spastica
spastica

Reputation: 33

How to extrapolate/interpolate within a dataset in Python

Assume I have a dataset below (This is a cut down version of the daily Treasury yields sourced from the Treasury website). I had artificially created some zeroes in there to illustrate the question I have. Each row corresponds to a given date.

Question - Assume I wanted to do log or quadratic extrapolation/interpolation within each row. Is there a quick way to do it or would one have to iterate through each row to fill it in?

    x = np.array([[ 1.31,  1.44,  0, 2.51,  0,  0],
                  [ 0,  1.45,  1.63, 2.48,  0,  2.69],
                  [ 1.31,  1.43,  1.59, 2.48,  2.55,  2.71]])

Upvotes: 0

Views: 148

Answers (1)

jrjames83
jrjames83

Reputation: 901

Here's a basic approach using the mean to interpolate the value. Of course you can apply a more formal calculation, but this is a start

import numpy as np

# Interpolate using the mean of each row

# Your original data
x = np.array([[ 1.31,  1.44,  0, 2.51,  0,  0],
              [ 0,  1.45,  1.63, 2.48,  0,  2.69],
              [ 1.31,  1.43,  1.59, 2.48,  2.55,  2.71]])

print(x)

print()

# for each row in x, set the zero values to the mean of the non zero values
for row in x:
    row[row == 0] = np.mean(row[np.nonzero(row)])

print(x)

Output below:

[[ 1.31  1.44  0.    2.51  0.    0.  ]
 [ 0.    1.45  1.63  2.48  0.    2.69]
 [ 1.31  1.43  1.59  2.48  2.55  2.71]]

[[ 1.31        1.44        1.75333333  2.51        1.75333333  1.75333333]
 [ 2.0625      1.45        1.63        2.48        2.0625      2.69      ]
 [ 1.31        1.43        1.59        2.48        2.55        2.71      ]]

Upvotes: 1

Related Questions