def corr_full(df):
ast.corr(df,
corr_method='pearson',
annot=True, palette='colorblind',
style='fivethirtyeight',
dpi=240,
title='This is a title',
sub_title='And this a subtitle',
x_label='this is x label',
y_label='and this y',
legend=False,
x_scale='log',
y_scale='symlog',
x_limit=1,
y_limit=[1,199])
def test_simple_minimal(df):
ast.corr(df)
ast.kde(data=df, x='A')
ast.hist(df, x='A')
ast.pie(df, x='other')
ast.swarm(df, x='A', y='B', hue='even')
ast.scat(df, x='A', y='B', hue='even')
ast.line(df, x='A')
ast.grid(df, x='A', y='B', col='even')
ast.box(df, x='odd', y='A', hue='even')
ast.violin(df, x='odd', y='A', hue='even')
ast.strip(df, x='odd', y='B', hue='even')
ast.count(df, x='cats')
ast.bargrid(df, x='even', y='B', hue='other', col='odd')
ast.overlap(df, x='A', y='B', label_col='other')
ast.multikde(df, x='A', label_col='even')
ast.compare(df, x='A', y=['B', 'C'], label_col='other')
ast.multicount(df, x='even', hue='odd', col='other')
def plot_corr(self, metric='val_acc', color_grades=5):
'''A heatmap with a single metric and hyperparameters.
NOTE: remember to invoke %matplotlib inline if in notebook
metric :: the metric to correlate against
color_grades :: number of colors to use in heatmap'''
cols = self._cols(metric)
return corr(self.data[cols], color_grades=color_grades)
def check_correlation(newData):
ast.corr(newData,
corr_method='pearson',
title="Correlation Heatmap of All Features")
plt.show()
contingency_table = np.zeros((3, 3), dtype=int)
contingency_table[0, 0] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 1))
contingency_table[0, 1] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 0))
contingency_table[0, 2] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 0.5))
contingency_table[1, 0] = sum(
(newData['Vaccinated'] == 0) == (newData['Dewormed'] == 1))
contingency_table[1, 1] = sum(
(newData['Vaccinated'] == 0) == (newData['Dewormed'] == 0))
contingency_table[1, 2] = sum(
(newData['Vaccinated'] == 0) == (newData['Dewormed'] == 0.5))
contingency_table[2, 0] = sum(
(newData['Vaccinated'] == 0.5) == (newData['Dewormed'] == 1))
contingency_table[2, 1] = sum(
(newData['Vaccinated'] == 0.5) == (newData['Dewormed'] == 0))
contingency_table[2, 2] = sum(
(newData['Vaccinated'] == 0.5) == (newData['Dewormed'] == 0.5))
stat, p, dof, expected = chi2_contingency(contingency_table)
prob = 0.95
# interpret p-value
alpha = 1.0 - prob
if p <= alpha:
print('Dependent (reject H0)')
else:
print('Independent (fail to reject H0)')
def plot_corr(self, metric, exclude, color_grades=5):
'''A heatmap with a single metric and hyperparameters.
NOTE: remember to invoke %matplotlib inline if in notebook
metric | str | Column label for the metric to correlate with
exclude | list | Column label/s to be excluded from the correlation
color_grades | int | Number of colors to use in heatmap
'''
try:
import astetik as ast
cols = self._cols(metric, exclude)
return ast.corr(self.data[cols], color_grades=color_grades)
except RuntimeError:
print('Matplotlib Runtime Error. Plots will not work.')
import pandas as pd
import matplotlib.pyplot as plt
from scipy.stats import chi2_contingency
newData = np.load("data_without_outliers.npy", allow_pickle=True)
newData = pd.DataFrame(newData, dtype=float)
newData.columns = [
'Type', 'Age', 'Breed1', 'Breed2', 'Gender', 'Color1', 'Color2', 'Color3',
'Maturity Size', 'Fur Length', 'Vaccinated', 'Dewormed', 'Sterilized',
'Health', 'Quantity', 'Fee', 'State', 'Video Amount', 'Photo Amount',
'Sentmt Magnitude', 'Sentmt Score', 'Adoption Speed'
]
ast.corr(newData,
corr_method='pearson',
title="Correlation Heatmap of All Features")
plt.show()
contingency_table = np.zeros((3, 3), dtype=int)
contingency_table[0, 0] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 1))
contingency_table[0, 1] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 0))
contingency_table[0, 2] = sum(
(newData['Vaccinated'] == 1) == (newData['Dewormed'] == 0.5))
contingency_table[1, 0] = sum(
(newData['Vaccinated'] == 0) == (newData['Dewormed'] == 1))
contingency_table[1, 1] = sum(
(newData['Vaccinated'] == 0) == (newData['Dewormed'] == 0))
