import warnings
warnings.filterwarnings("ignore")

import pandas as pd
import matplotlib.pyplot as plt

# Download and load shampoo sales data
url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/shampoo.csv"
shampoo = pd.read_csv(url)
print("Shape:", shampoo.shape)
print(shampoo.head())

# Check column names and types
print(shampoo.dtypes)

shampoo['Month'] = pd.date_range(start='1901-01', periods=shampoo.shape[0], freq='M')
shampoo['Month'] = pd.to_datetime(shampoo['Month'])

# Line plot to visualize sales over time
plt.figure(figsize=(8,4))
plt.plot(shampoo['Month'], shampoo['Sales'])
plt.title("Monthly Shampoo Sales")
plt.xlabel("Month")
plt.ylabel("Sales")
plt.show()

# Simple scatter plot: Month vs. Sales
plt.figure(figsize=(8,4))
plt.scatter(shampoo['Month'], shampoo['Sales'], color='teal')
plt.title("Scatter Plot: Shampoo Sales Over Time")
plt.xlabel("Month")
plt.ylabel("Sales")
plt.show()

# Creating a synthetic example: Advertising vs. Sales
import numpy as np
np.random.seed(0)
shampoo['Advertising'] = np.random.randint(200, 400, size=shampoo.shape[0])
# Scatter plot of Advertising vs Sales
plt.figure(figsize=(8,4))
plt.scatter(shampoo['Advertising'], shampoo['Sales'], c='green')
plt.title("Sales vs. Advertising")
plt.xlabel("Advertising Spend")
plt.ylabel("Sales")
plt.show()

# Show Histogram of Sales data
plt.figure(figsize=(8,4))
plt.hist(shampoo['Sales'], bins=8, color='skyblue', edgecolor='black')
plt.title("Distribution of Monthly Sales")
plt.xlabel("Sales")
plt.ylabel("Number of Months")
plt.show()

# Try with fewer or more bins
plt.figure(figsize=(8,4))
plt.hist(shampoo['Sales'], bins=3, color='salmon', edgecolor='white', alpha=0.8)
plt.title("Histogram with 3 Bins")
plt.xlabel("Sales")
plt.ylabel("Count")
plt.show()

# Compare two different columns in one histogram
plt.figure(figsize=(8,4))
plt.hist([shampoo['Sales'], shampoo['Advertising']], bins=8, alpha=0.7, label=['Sales','Advertising'])
plt.title("Sales vs. Advertising Distributions")
plt.xlabel("Value")
plt.ylabel("Number of Months")
plt.legend()
plt.show()

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

# Drop missing values just in case
shampoo_clean = shampoo.dropna()

# Mini-project: Find months with top 5 highest sales
top_months = shampoo_clean.sort_values(by='Sales', ascending=False).head(5)
print(top_months[['Month', 'Sales']])

# Visualize top 5 sales months on a scatter plot
plt.figure(figsize=(8,4))
plt.scatter(shampoo_clean['Month'], shampoo_clean['Sales'], label='All Months', alpha=0.5)
plt.scatter(top_months['Month'], top_months['Sales'], color='red', label='Top 5 Months', s=80)
plt.title("Sales: Top 5 Months Highlighted")
plt.xlabel("Month")
plt.ylabel("Sales")
plt.legend()
plt.show()

# Practice: Try your own histogram with user input
import builtins
column = input("Type 'Sales' or 'Advertising': ")
if column in shampoo_clean.columns:
    plt.figure(figsize=(8,4))
    plt.hist(shampoo_clean[column], bins=6, color='gold', edgecolor='black')
    plt.title(f"Histogram of {column}")
    plt.xlabel(column)
    plt.ylabel("Count")
    plt.show()
else:
    print("Column not found.")
    

# Challenge: Make a scatter plot with different dot sizes
sizes = (shampoo_clean['Advertising'] - shampoo_clean['Advertising'].min()) * 0.8
plt.figure(figsize=(8,4))
plt.scatter(shampoo_clean['Month'], shampoo_clean['Sales'], s=sizes, alpha=0.6, color='purple')
plt.title("Sales with Dot Sizes by Advertising")
plt.xlabel("Month")
plt.ylabel("Sales")
plt.show()