import pandas as pd
from pandas import DataFrame
import numpy as np
from sklearn.decomposition import PCA
Create a DataFrame full of random numbers.
df = pd.DataFrame(data=np.random.normal(0, 1, (50, 8)))
df.head()
| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | |
|---|---|---|---|---|---|---|---|---|
| 0 | -0.314008 | -1.090192 | -0.920020 | -0.559545 | 0.775226 | -0.882474 | 1.050049 | -0.092116 |
| 1 | -0.145504 | -0.297851 | 0.140186 | -0.724637 | 0.558747 | -0.708013 | -0.555629 | 0.653227 |
| 2 | 0.711590 | -0.215502 | -0.266352 | -1.058049 | -0.324418 | 0.929751 | 0.352552 | -0.256500 |
| 3 | 0.625027 | -0.725645 | 0.477974 | -1.182262 | 1.435128 | -0.673370 | -1.112910 | 0.354148 |
| 4 | 1.335500 | -0.488708 | -0.131384 | -0.581381 | -0.451392 | 0.149943 | -0.250088 | -0.011798 |
Fit the PCA.
pca = PCA(n_components=4)
pca.fit(df)
PCA(copy=True, iterated_power='auto', n_components=4, random_state=None,
svd_solver='auto', tol=0.0, whiten=False)
Apply the transformation and convert the result into a DataFrame.
columns = ['pca_%i' % i for i in range(4)]
df_pca = DataFrame(pca.transform(df), columns=columns, index=df.index)
df_pca.head()
| pca_0 | pca_1 | pca_2 | pca_3 | |
|---|---|---|---|---|
| 0 | -0.546917 | 0.506951 | -1.042971 | 0.287119 |
| 1 | -0.054910 | -0.217007 | -0.539264 | 0.457777 |
| 2 | 0.908084 | -0.113016 | -0.602685 | -1.009475 |
| 3 | -0.168224 | -1.419395 | -1.216016 | 0.196382 |
| 4 | 0.678731 | -0.605830 | -0.057639 | -0.428901 |
If you spot any errors please let me know on twitter.