Hypothesis Testing in Python

This project demonstrates how to apply parametric and non-parametric hypothesis tests using a real-world Car Dataset.
It covers tests like t-test, ANOVA, Chi-Square, Mann–Whitney, and Kruskal–Wallis with Python implementations, visualizations, and short interpretations.

---

📊 Dataset Overview

The dataset contains information on 428 cars, including attributes such as:

• Make, Model, Type, Origin, DriveTrain
• Engine Size, Cylinders, Horsepower, MPG (City & Highway)
• MSRP, Invoice, Weight, Wheelbase, Length

Used to explore relationships between numerical and categorical variables.

---

🧠 Key Objectives

• Perform parametric and non-parametric hypothesis testing
• Compare mean differences across groups (Origin, Type, DriveTrain)
• Check normality and variance equality
• Visualize and interpret results for practical decision-making

---

⚙️ Tests Included

Test	Use Case	Example Hypothesis	Decision Basis
One-Sample t-test	Compare mean to known value	Horsepower ≠ 200	Based on p-value
Independent t-test	Compare two origins	USA vs Europe MSRP	Based on p-value
ANOVA	Compare 3+ groups	Sedan vs SUV vs Sports	Based on p-value
Mann–Whitney	Non-parametric 2 groups	USA vs Europe	Based on p-value
Kruskal–Wallis	Non-parametric ANOVA	Car Types	Based on p-value
Chi-Square	Categorical relationship	Origin vs DriveTrain	Based on p-value

---

📈 Example Code

from scipy.stats import ttest_ind, f_oneway, chi2_contingency

# Example: Compare MSRP by Origin (Independent t-test)
usa = df[df['Origin'] == 'USA']['MSRP']
europe = df[df['Origin'] == 'Europe']['MSRP']

t_stat, p_val = ttest_ind(usa, europe)
print("T-statistic:", t_stat, "P-value:", p_val)
if p_val < 0.05:
    print("Reject H₀ → Significant difference in average price.")
else:
    print("Fail to reject H₀ → No significant difference.")


# Author

- Awais Manzoor
- Data Analyst | Python & SQL Enthusiast