import pandas as pd
import numpy as np
print(pd.__version__)

rng = np.random.default_rng(seed=99)

cuisines = ['Italian', 'Japanese', 'Mexican', 'Indian', 'Thai']
cities = ['Portside', 'Elmfield', 'Brookhaven']
price_ranges = ['Budget', 'Mid-Range', 'Upscale']

restaurants = pd.DataFrame({
    'restaurant_id': range(1, 26),
    'restaurant_name': [f'Restaurant {i}' for i in range(1, 26)],
    'cuisine': rng.choice(cuisines, size=25),
    'city': rng.choice(cities, size=25),
    'price_range': rng.choice(price_ranges, size=25, p=[0.35, 0.45, 0.2])
})
restaurants.head()

n_reviews = 600
review_rows = []
review_dates = pd.date_range('2026-01-01', periods=210, freq='D')

for review_id in range(1, n_reviews + 1):
    restaurant_id = int(rng.choice(restaurants['restaurant_id']))
    rating = int(rng.choice([1, 2, 3, 4, 5], p=[0.05, 0.1, 0.2, 0.35, 0.3]))
    review_date = rng.choice(review_dates)
    review_rows.append([review_id, restaurant_id, rating, review_date])

reviews = pd.DataFrame(review_rows, columns=['review_id', 'restaurant_id', 'rating', 'review_date'])
print(reviews.shape)

duplicate_rows = reviews.sample(n=20, random_state=5)
reviews = pd.concat([reviews, duplicate_rows], ignore_index=True)
reviews.to_csv('restaurant_reviews.csv', index=False)
reviews = pd.read_csv('restaurant_reviews.csv', parse_dates=['review_date'])
print(reviews.shape)

print(reviews.duplicated().sum())
reviews[reviews.duplicated()].head()

reviews = reviews.drop_duplicates().reset_index(drop=True)
print(reviews.shape)
print(reviews.duplicated().sum())

full_reviews = reviews.merge(restaurants, on='restaurant_id', how='left')
full_reviews[['restaurant_name', 'cuisine', 'city', 'rating']].head()

restaurant_stats = full_reviews.groupby('restaurant_name').agg(
    review_count=('rating', 'count'),
    avg_rating=('rating', 'mean')
).round(2)
top_rated = restaurant_stats[restaurant_stats['review_count'] >= 10].sort_values('avg_rating', ascending=False)
top_rated.head(5)

cuisine_summary = full_reviews.groupby('cuisine').agg(
    review_count=('rating', 'count'),
    avg_rating=('rating', 'mean')
).round(2).sort_values('avg_rating', ascending=False)
cuisine_summary

full_reviews['rating'].value_counts().sort_index()

cuisine_price_pivot = pd.pivot_table(full_reviews, values='rating', index='cuisine', columns='price_range', aggfunc='mean').round(2)
cuisine_price_pivot = cuisine_price_pivot[['Budget', 'Mid-Range', 'Upscale']]
cuisine_price_pivot

pd.crosstab(full_reviews['cuisine'], full_reviews['city'])

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(8, 5))
cuisine_summary['avg_rating'].plot(kind='bar', ax=ax, color='darkorange', ylim=(0, 5))
ax.set_title('Average Rating by Cuisine')
ax.set_ylabel('Average Rating (out of 5)')
plt.tight_layout()
plt.savefig('avg_rating_by_cuisine.png', dpi=150)
plt.close(fig)
print('Saved avg_rating_by_cuisine.png')

fig, ax = plt.subplots(figsize=(8, 5))
cuisine_price_pivot.plot(kind='bar', ax=ax, ylim=(0, 5))
ax.set_title('Average Rating by Cuisine and Price Range')
ax.set_ylabel('Average Rating (out of 5)')
ax.legend(title='Price Range')
plt.tight_layout()
plt.savefig('rating_by_cuisine_and_price.png', dpi=150)
plt.close(fig)
print('Saved rating_by_cuisine_and_price.png')