License


Copyright (C) 2017 J. Patrick Hall, jphall@gwu.edu

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.


Simple standardization - Pandas and numpy

Imports


In [2]:
import pandas as pd              # pandas for handling mixed data sets 
import numpy as np               # numpy for basic math and matrix operations

Create sample data set


In [3]:
# create a data frame containing variables of disparate scale
scratch_df = pd.DataFrame({'x1': pd.Series(np.random.choice(1000, 20)),
                           'x2': pd.Series(np.random.choice(20, 20))}) 

scratch_df


Out[3]:
x1 x2
0 68 18
1 106 3
2 878 2
3 791 7
4 812 12
5 348 13
6 209 10
7 818 17
8 525 9
9 125 0
10 423 3
11 411 5
12 859 6
13 908 18
14 953 16
15 567 11
16 105 4
17 590 18
18 145 2
19 425 9

Standardize


In [6]:
# create a deep copy 
# so this cell can be run many times w/o error
scratch_df1 = scratch_df.copy()

# loop through columns
# create new column
# apply z-score formula to new column
for col_name in scratch_df.columns:
    new_col_name = col_name + '_std'
    scratch_df1[new_col_name] = (scratch_df[col_name] - scratch_df[col_name].mean())/scratch_df[col_name].std()

# new variables are on the same scale
scratch_df1


Out[6]:
x1 x2 x1_std x2_std
0 68 18 -1.402092 1.467988
1 106 3 -1.279695 -1.020127
2 878 2 1.206901 -1.186002
3 791 7 0.926675 -0.356630
4 812 12 0.994316 0.472742
5 348 13 -0.500218 0.638616
6 209 10 -0.947934 0.140993
7 818 17 1.013642 1.302114
8 525 9 0.069895 -0.024881
9 125 0 -1.218496 -1.517750
10 423 3 -0.258645 -1.020127
11 411 5 -0.297296 -0.688379
12 859 6 1.145702 -0.522504
13 908 18 1.303530 1.467988
14 953 16 1.448474 1.136239
15 567 11 0.205176 0.306868
16 105 4 -1.282916 -0.854253
17 590 18 0.279259 1.467988
18 145 2 -1.154076 -1.186002
19 425 9 -0.252203 -0.024881