9/28/24, 3:14 PM Python Session 14092024.

ipynb - Colab

# Import necessary libraries

import numpy as np
import scipy.stats as stats
import pandas as pd
from statsmodels.stats.weightstats import ztest

Z-Test

keyboard_arrow_down Scenario 1: Testing if the average weight of newborns differs from the hospital's claim
# 1. Z-Test
# Scenario 1: Testing if the average weight of newborns differs from the hospital's claim
sample_mean = 3.4
population_mean = 3.5
std_dev = 0.5
n = 50
z_score = (sample_mean - population_mean) / (std_dev / np.sqrt(n))
p_value = stats.norm.cdf(z_score) * 2 # Two-tailed test
print("Z-Test (Scenario 1):")
print(f"Z-Score: {z_score}")
print(f"P-Value: {p_value}\n")

Z-Test (Scenario 1):

Z-Score: -1.4142135623730963
P-Value: 0.1572992070502848

Start coding or generate with AI.

keyboard_arrow_down Scenario 2: Testing if a marketing campaign increased average daily sales

sample_mean = 2100
population_mean = 2000
std_dev = 300
n = 50
z_score = (sample_mean - population_mean) / (std_dev / np.sqrt(n))
p_value = stats.norm.cdf(z_score) * 2 # Two-tailed test
print("Z-Test (Scenario 2):")
print(f"Z-Score: {z_score}")
print(f"P-Value: {p_value}\n")

Start coding or generate with AI.

Start coding or generate with AI.

t-test

One-Sample T-Test

keyboard_arrow_down Scenario 1: Testing if the average score of students differs from 75

# Scenario 1: Testing if the average score of students differs from 75
sample_scores = np.array([70, 72, 68, 74, 76, 69, 71, 73, 70, 75, 77, 74, 73, 69, 72])
t_score, p_value = stats.ttest_1samp(sample_scores, 75)
print("One-Sample T-Test (Scenario 1):")
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 1/6
9/28/24, 3:14 PM Python Session 14092024.ipynb - Colab

keyboard_arrow_down Scenario 2: Testing if a new sales strategy increased monthly sales from $5000
# Scenario 2: Testing if a new sales strategy increased monthly sales from $5000
sample_sales = np.array([5200, 5100, 5300, 5400, 5000, 5100, 5200, 5300, 5000, 5400,
5200, 5300, 5100, 5200, 5400, 5300, 5100, 5200, 5300, 5400])
t_score, p_value = stats.ttest_1samp(sample_sales, 5000)
print("One-Sample T-Test (Scenario 2):")
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

Start coding or generate with AI.

Two-Sample T-Test

keyboard_arrow_down Scenario 1: Testing if there is a difference in average test scores between two classes
# Scenario 1: Testing if there is a difference in average test scores between two classes
class_A_scores = np.array([80, 82, 85, 87, 79, 83, 86, 88, 82, 84, 85, 86, 89, 81, 87])
class_B_scores = np.array([75, 77, 79, 76, 78, 77, 76, 79, 78, 77, 75, 76, 77, 78, 76])
t_score, p_value = stats.ttest_ind(class_A_scores, class_B_scores)
print("Two-Sample T-Test (Scenario 1):")
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

Start coding or generate with AI.

keyboard_arrow_down Scenario 2: Testing if there is a difference in sales between two stores

# Scenario 2: Testing if there is a difference in sales between two stores
store_A_sales = np.array([3000, 3100, 3200, 3300, 3400, 3100, 3200, 3000, 3100, 3200,
3300, 3400, 3000, 3100, 3200, 3000, 3100, 3300, 3400, 3100,
3200, 3000, 3100, 3200, 3300, 3400, 3000, 3100, 3200, 3300])
store_B_sales = np.array([3200, 3300, 3400, 3200, 3300, 3200, 3400, 3300, 3400, 3200,
3300, 3400, 3200, 3300, 3400, 3200, 3300, 3200, 3400, 3300,
3400, 3200, 3300, 3400, 3200, 3300, 3400, 3200, 3300, 3400])
t_score, p_value = stats.ttest_ind(store_A_sales, store_B_sales)
print("Two-Sample T-Test (Scenario 2):")
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

keyboard_arrow_down Paired T-Test

# Scenario 1: Testing if a training program improves productivity
before_training = np.array([10, 12, 11, 13, 12, 11, 14, 12, 15, 13])
after_training = np.array([15, 16, 15, 18, 17, 16, 19, 17, 20, 18])
t_score, p_value = stats.ttest_rel(before_training, after_training)
print("Paired T-Test (Scenario 1):")
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

# Scenario 2: Testing if a promotional offer impacts sales

before_promo = np.array([5000, 5100, 5200, 5300, 5000, 5100, 5200, 5300, 5000, 5400,
5200, 5300, 5100, 5200, 5400])
after_promo = np.array([5300, 5400, 5500, 5600, 5300, 5400, 5500, 5600, 5300, 5700,
5500, 5600, 5400, 5500, 5700])
t_score, p_value = stats.ttest_rel(before_promo, after_promo)
print("Paired T-Test (Scenario 2):")

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 2/6
9/28/24, 3:14 PM Python Session 14092024.ipynb - Colab
print(f"T-Score: {t_score}")
print(f"P-Value: {p_value}\n")

Paired T-Test (Scenario 1):

T-Score: -36.0
P-Value: 4.873628935030659e-11

Paired T-Test (Scenario 2):

T-Score: -inf
P-Value: 0.0

C:\Users\admin\AppData\Local\Temp\ipykernel_5468\1620158596.py:14: RuntimeWarning: Precision loss occurred in moment calculation due

t_score, p_value = stats.ttest_rel(before_promo, after_promo)

Start coding or generate with AI.

keyboard_arrow_down 1. One-Way ANOVA

# Dataset for One-Way ANOVA (Advertising Budget vs. Sales for 3 different campaigns)
data = {'Campaign 1': [500, 700, 800, 900, 950],
'Campaign 2': [1000, 1100, 1200, 1250, 1300],
'Campaign 3': [400, 450, 500, 550, 600]}

df = pd.DataFrame(data)

df

Campaign 1 Campaign 2 Campaign 3

0 500 1000 400

1 700 1100 450

2 800 1200 500

3 900 1250 550

4 950 1300 600

# One-Way ANOVA
f_stat, p_val = stats.f_oneway(df['Campaign 1'], df['Campaign 2'], df['Campaign 3'])

print("F-Statistic:", f_stat)
print("P-Value:", p_val)

F-Statistic: 32.31279620853076
P-Value: 1.4751779256171162e-05

# Since p value is less than 0.5, hence we will reject the null hypothesis,
# Hence there is not statistical difference between these 3 campaigns

Start coding or generate with AI.

keyboard_arrow_down 2. Two-Way ANOVA

import statsmodels.api as sm
from statsmodels.formula.api import ols

# Dataset for Two-Way ANOVA

data = {'Sales': [500, 700, 800, 950, 400, 600, 1000, 1200, 1300, 1250],
'Ad_Campaign': ['Campaign 1', 'Campaign 1', 'Campaign 1', 'Campaign 1', 'Campaign 2',
'Campaign 2', 'Campaign 2', 'Campaign 2', 'Campaign 3', 'Campaign 3'],
'Store': ['Store A', 'Store B', 'Store A', 'Store B', 'Store A',
'Store B', 'Store A', 'Store B', 'Store A', 'Store B']}

df = pd.DataFrame(data)

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 3/6
9/28/24, 3:14 PM Python Session 14092024.ipynb - Colab
# Two-Way ANOVA
model = ols('Sales ~ C(Ad_Campaign) + C(Store) + C(Ad_Campaign):C(Store)', data=df).fit()
anova_table = sm.stats.anova_lm(model, typ=2)
print(anova_table)

sum_sq df F PR(>F)
C(Ad_Campaign) 417875.0 2.0 1.915759 0.260872
C(Store) 49000.0 1.0 0.449284 0.539387
C(Ad_Campaign):C(Store) 22875.0 2.0 0.104871 0.902836
Residual 436250.0 4.0 NaN NaN

Start coding or generate with AI.

keyboard_arrow_down 3. Pearson’s Correlation

import pandas as pd
from scipy.stats import pearsonr

# Dataset (Advertising Budget vs. Sales)

data = {'Advertising Budget ($)': [10000, 20000, 15000, 25000, 30000],
'Sales (Units)': [500, 1200, 850, 1500, 2000]}

df = pd.DataFrame(data)

# Pearson's Correlation
corr, _ = pearsonr(df['Advertising Budget ($)'], df['Sales (Units)'])
print(f'Pearson Correlation: {corr}')

Pearson Correlation: 0.9963607400395207

keyboard_arrow_down 4. Spearman’s Rank Correlation

from scipy.stats import spearmanr

# Dataset (Customer Satisfaction Rank vs. Revenue Rank)

data = {'Customer Satisfaction Rank': [1, 3, 2, 4],
'Revenue Rank': [2, 4, 1, 3]}

df = pd.DataFrame(data)

# Spearman's Correlation
corr, _ = spearmanr(df['Customer Satisfaction Rank'], df['Revenue Rank'])
print(f'Spearman Correlation: {corr}')

Spearman Correlation: 0.6000000000000001

keyboard_arrow_down 5. Point-Biserial Correlation

from scipy.stats import pointbiserialr

# Dataset (Coupon Usage vs. Amount Spent)

data = {'Used Coupon': [1, 0, 1, 0], # Binary variable: 1 = Yes, 0 = No
'Amount Spent ($)': [250, 180, 300, 120]}

df = pd.DataFrame(data)

# Point-Biserial Correlation
corr, _ = pointbiserialr(df['Used Coupon'], df['Amount Spent ($)'])
print(f'Point-Biserial Correlation: {corr}')

Point-Biserial Correlation: 0.9147021665807529

keyboard_arrow_down 6. Cross-Correlation (Time-Series)

import numpy as np
import matplotlib.pyplot as plt

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 4/6
9/28/24, 3:14 PM Python Session 14092024.ipynb - Colab

# Time-series dataset (Social Media Mentions vs. Sales)

social_media_mentions = np.array([500, 600, 700, 850])
sales = np.array([50000, 52000, 55000, 70000])

# Cross-Correlation
cross_corr = np.correlate(social_media_mentions - np.mean(social_media_mentions), sales - np.mean(sales), mode='full')
lags = np.arange(-len(sales) + 1, len(sales))

# Plot the cross-correlation

plt.plot(lags, cross_corr)
plt.title('Cross-Correlation between Social Media Mentions and Sales')
plt.xlabel('Lag')
plt.ylabel('Cross-Correlation')
plt.show()

keyboard_arrow_down 7. Kendall’s Tau Correlation

from scipy.stats import kendalltau

# Dataset (Performance Rank vs. Promotion Rank)

data = {'Performance Rank': [2, 1, 3, 2],
'Promotion Rank': [3, 1, 4, 2]}

df = pd.DataFrame(data)

# Kendall's Tau Correlation

corr, _ = kendalltau(df['Performance Rank'], df['Promotion Rank'])
print(f'Kendall Tau Correlation: {corr}')

Kendall Tau Correlation: 0.912870929175277

Start coding or generate with AI.

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 5/6
9/28/24, 3:14 PM Python Session 14092024.ipynb - Colab

https://colab.research.google.com/drive/1Vp85nKWtvXIv2ZReKpIguicdBscWpeVv#printMode=true 6/6