# Suppress warnings for a cleaner notebook
import warnings; warnings.filterwarnings("ignore")

# Import key packages
import pandas as pd
import numpy as np
from scipy import stats
import seaborn as sns
import matplotlib.pyplot as plt

# Data setup
tips = sns.load_dataset("tips")
print("Shape:", tips.shape)
tips.head()

# Quick explore: average tip on weekend vs weekday
tips["day_type"] = tips["day"].apply(lambda x: "Weekend" if x in ["Sat", "Sun"] else "Weekday")
weekend_mean = tips.query('day_type == "Weekend"')["tip"].mean()
weekday_mean = tips.query('day_type == "Weekday"')["tip"].mean()
print("Average tip on weekends:", round(weekend_mean, 2))
print("Average tip on weekdays:", round(weekday_mean, 2))

# Doing the independent t-Test
weekend_tips = tips.query('day_type == "Weekend"')["tip"]
weekday_tips = tips.query('day_type == "Weekday"')["tip"]
t_stat, p_val = stats.ttest_ind(weekend_tips, weekday_tips, equal_var=False)
print("t-statistic:", round(t_stat, 3))
print("p-value:", p_val)

# Visualize the difference in average tips
sns.boxplot(x="day_type", y="tip", data=tips, palette="Set2");
plt.title("Tip Distribution: Weekend vs Weekday")
plt.ylabel("Tip Amount ($)")
plt.show()

# One-way ANOVA: average tips by each day
days = [day for day in tips["day"].unique()]
groups = [tips.query('day == @d')["tip"] for d in days]
anova_stat, anova_p = stats.f_oneway(*groups)
print("ANOVA F-statistic:", round(anova_stat, 3))
print("ANOVA p-value:", anova_p)

# Visualize tips for each day
sns.boxplot(x="day", y="tip", data=tips, palette="Set3");
plt.title("Tip Distribution by Day")
plt.ylabel("Tip Amount ($)")
plt.show()

# Chi-Square test: Are males/females equally likely to tip above $3?
tips["tip_above_3"] = tips["tip"] > 3
contingency = pd.crosstab(tips["sex"], tips["tip_above_3"])
chi2, p, dof, ex = stats.chi2_contingency(contingency)
print("Chi-square statistic:", round(chi2, 3))
print("p-value:", p)
contingency

# Visualize the rate of tips above $3 by gender
tip_rates = contingency.div(contingency.sum(axis=1), axis=0)
tip_rates[True].plot(kind="bar", color=["skyblue", "salmon"]);
plt.ylabel("Fraction tipping above $3")
plt.title("Tip above $3: Male vs Female")
plt.ylim(0,1)
plt.show()

# Let us confirm: What is our sample size for the main tests?
print("Number of rows:", len(tips))
tips["sex"].value_counts()

# Check for missing data
print(tips.isnull().sum())

# Drop rows with missing values for our columns of interest
clean_tips = tips.dropna(subset=["tip", "day", "sex"]).copy()
print("Rows after cleaning:", len(clean_tips))

# Solution: t-Test for smokers vs non-smokers
smoker_tips = clean_tips[clean_tips["smoker"]=="Yes"]["tip"]
nonsmoker_tips = clean_tips[clean_tips["smoker"]=="No"]["tip"]
t_stat_smoker, p_val_smoker = stats.ttest_ind(smoker_tips, nonsmoker_tips, equal_var=False)
print("Smoker avg:", round(smoker_tips.mean(),2))
print("Non-smoker avg:", round(nonsmoker_tips.mean(),2))
print("t-statistic:", round(t_stat_smoker, 3))
print("p-value:", p_val_smoker)

# Ask user: Would you expect smokers or non-smokers to tip more? Why?
user_guess = input("Who do you think tips more on average: smokers or non-smokers? Why? ")
print("You guessed:", user_guess)