# Suppress warnings to keep outputs clean
import warnings
from statsmodels.tools.sm_exceptions import ConvergenceWarning, ValueWarning
warnings.filterwarnings("ignore", category=ValueWarning)
warnings.filterwarnings("ignore", category=ConvergenceWarning)
warnings.filterwarnings("ignore", category=RuntimeWarning)
import warnings as _w; _w.filterwarnings("ignore")  # Extra line for broad suppression

# Let's simulate flipping a coin 10 times
import numpy as np
coin_flips = np.random.choice(['Heads', 'Tails'], size=10)
print("Coin flips:", coin_flips)

dice_rolls = np.random.choice(np.arange(1, 7), size=20)
print("Dice rolls:", dice_rolls)

import scipy.stats as stats
mean_roll = np.mean(dice_rolls)
median_roll = np.median(dice_rolls)
mode_result = stats.mode(dice_rolls, keepdims=True)
mode_roll = mode_result.mode[0] if hasattr(mode_result, "mode") else mode_result[0]
print("Mean:", mean_roll)
print("Median:", median_roll)
print("Mode:", mode_roll)

# Data setup
import pandas as pd
import seaborn as sns
titanic = sns.load_dataset("titanic")
print("Shape:", titanic.shape)
print(titanic.head())

# Check for missing values
print(titanic.isnull().sum())

# Fill missing age values with the median age
titanic['age'] = titanic['age'].fillna(titanic['age'].median())
# Drop any rows missing 'embarked'
titanic = titanic.dropna(subset=['embarked'])

# Let us look at the chance of survival
survival_rate = titanic['survived'].mean()
print("Survival rate:", survival_rate)

# Simulating 1000 coin flips
flips = np.random.binomial(n=1, p=0.5, size=1000)
print("Number of heads:", np.sum(flips))

import matplotlib.pyplot as plt
normal_data = np.random.normal(loc=0, scale=1, size=500)
plt.hist(normal_data, bins=30, alpha=0.7, color='skyblue')
plt.title("Normal Distribution Example")
plt.show()

# Exponential distribution demo
exp_data = np.random.exponential(scale=2.0, size=500)
plt.hist(exp_data, bins=30, color='salmon', alpha=0.8)
plt.title("Exponential Distribution")
plt.show()

# Take 100 samples of 20 ages each, find the mean each time
sample_means = [titanic['age'].sample(20, replace=True).mean() for _ in range(100)]
plt.hist(sample_means, bins=20, color='mediumpurple', alpha=0.8)
plt.title("Distribution of Sample Means (CLT Example)")
plt.xlabel("Sample Mean Age")
plt.show()

from scipy.stats import sem, t
ages = titanic['age'].dropna()
mean_age = ages.mean()
ci = t.interval(0.95, len(ages)-1, loc=mean_age, scale=sem(ages))
print(f"95% confidence interval for mean age: {ci}")

survived_f = titanic[titanic['sex']=='female']['survived']
survived_m = titanic[titanic['sex']=='male']['survived']
from scipy.stats import ttest_ind
t_stat, p_value = ttest_ind(survived_f, survived_m, nan_policy="omit")
print(f"p-value: {p_value}")

corr = titanic['age'].corr(titanic['fare'])
covar = titanic['age'].cov(titanic['fare'])
print(f"Correlation: {corr}")
print(f"Covariance: {covar}")

from sklearn.linear_model import LinearRegression
X = titanic[['age']].values
y = titanic['fare'].values
reg = LinearRegression()
reg.fit(X, y)
pred_fare = reg.predict([[30]])
print(f"Predicted fare for age 30: {pred_fare[0]:.2f}")

from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
df = titanic.copy()
df['sex'] = (df['sex'] == 'male').astype(int)
X = df[['age', 'sex', 'pclass']].values
y = df['survived'].values
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
clf = LogisticRegression(max_iter=200)
clf.fit(X_train, y_train)
score = clf.score(X_test, y_test)
print(f"Test set accuracy: {score:.2f}")

air_url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/airline-passengers.csv"
airline = pd.read_csv(air_url)
plt.plot(pd.to_datetime(airline['Month']), airline['Passengers'])
plt.title("Monthly Airline Passengers")
plt.xlabel("Month")
plt.ylabel("Passengers")
plt.show()

# Let's work through Bayes' theorem: spam emails
# Suppose 20% of emails are spam (P(Spam)=0.2).
# 95% of spam emails contain 'win' (P(Win|Spam)=0.95),
# and 5% of non-spam emails contain 'win' (P(Win|NotSpam)=0.05).
# If you see 'win', what's the chance it's spam?
P_spam = 0.2
P_win_given_spam = 0.95
P_win_given_notspam = 0.05
P_notspam = 1 - P_spam
P_win = P_win_given_spam*P_spam + P_win_given_notspam*P_notspam
P_spam_given_win = (P_win_given_spam*P_spam) / P_win
print(f"P(Spam | Win): {P_spam_given_win:.2f}")

N = 10000
ill = np.random.choice([1, 0], size=N, p=[0.01, 0.99])
test_pos = np.zeros(N)
for i in range(N):
    if ill[i]:
        test_pos[i] = np.random.rand() < 0.99  # True positive
    else:
        test_pos[i] = np.random.rand() < 0.05  # False positive
# Bayesian answer: P(Ill | Test+) = 
num_positive = np.sum(test_pos)
num_ill_positive = np.sum((ill == 1) & (test_pos == 1))
prob = num_ill_positive / num_positive
print(f"Chance you are really ill if positive: {prob:.3f}")

num_class3 = (titanic['pclass']==3).sum()
num_surv_class3 = ((titanic['pclass']==3) & (titanic['survived']==1)).sum()
prob_surv_class3 = num_surv_class3 / num_class3
print(f"Chance a 3rd-class passenger survived: {prob_surv_class3:.2f}")