import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')

# Load sample multimodal social media analytics data (posts, engagement, time series)
np.random.seed(42)

# Content dataset: Social Media Content (multiple platforms, posts)
n_posts = 500
views = np.random.randint(100, 100000, n_posts)
sm_content_df = pd.DataFrame({
    'post_id': range(1, n_posts+1),
    'platform': np.random.choice(['YouTube','Instagram','TikTok'], n_posts),
    'date': pd.date_range('2023-01-01', periods=n_posts, freq='6h'),
    'views': views,
    'likes': (views * np.random.uniform(0.02, 0.15, n_posts)).astype(int),
    'comments': (views * np.random.uniform(0.001, 0.05, n_posts)).astype(int),
    'shares': (views * np.random.uniform(0.001, 0.03, n_posts)).astype(int)
})
print(sm_content_df.head(3))

# Load and preview social_media_time_series (daily audience trends)
n_days = 365
dates = pd.date_range('2023-01-01', periods=n_days, freq='D')
views = np.random.randint(1000, 50000, n_days)
likes = (views * np.random.uniform(0.03, 0.12, n_days)).astype(int)
time_series_df = pd.DataFrame({
    'date': dates,
    'views': views,
    'likes': likes,
    'engagement_rate': np.round(likes / views * 100, 2)
})
print(time_series_df.head(3))

# Calculate engagement rate column
sm_content_df['engagement_rate'] = (sm_content_df['likes'] + sm_content_df['comments'] + sm_content_df['shares']) / sm_content_df['views'] * 100
print(sm_content_df[['post_id', 'views', 'likes', 'comments', 'shares', 'engagement_rate']].head(5))

# Calculate average engagement rate by platform
platform_means = sm_content_df.groupby('platform')['engagement_rate'].mean().sort_values(ascending=False)
print(platform_means)

# Find top 5 posts by engagement rate
top5_engaged = sm_content_df.sort_values('engagement_rate', ascending=False).head(5)
print(top5_engaged[['post_id', 'platform', 'views', 'engagement_rate']])

# Calculate monthly average engagement rate
sm_content_df['month'] = sm_content_df['date'].dt.to_period('M')
monthly_trends = sm_content_df.groupby('month')['engagement_rate'].mean()
print(monthly_trends.head())

# Plot engagement rate over time (line chart)
import matplotlib.pyplot as plt
monthly_trends.plot(title='Average Monthly Engagement Rate (%)', marker='o')
plt.ylabel('Engagement Rate (%)')
plt.xlabel('Month')
plt.tight_layout()
plt.show()

# Calculate audience growth: monthly total views
monthly_views = sm_content_df.groupby('month')['views'].sum()
monthly_likes = sm_content_df.groupby('month')['likes'].sum()
plt.plot(monthly_views.index.astype(str), monthly_views.values, label='Total Views')
plt.plot(monthly_likes.index.astype(str), monthly_likes.values, label='Total Likes')
plt.title('Monthly Audience Growth')
plt.xlabel('Month')
plt.ylabel('Count')
plt.legend()
plt.tight_layout()
plt.show()

# Define viral threshold: engagement rate > 2 std deviations above mean
mean_rate = sm_content_df['engagement_rate'].mean()
std_rate = sm_content_df['engagement_rate'].std()
viral_threshold = mean_rate + 2 * std_rate
viral_posts = sm_content_df[sm_content_df['engagement_rate'] > viral_threshold]
print('Viral Posts Found:', len(viral_posts))
print(viral_posts[['post_id', 'platform', 'views', 'engagement_rate']])

# Platform and time-based audience segment analysis
segment_trends = sm_content_df.groupby(['platform', 'month'])['engagement_rate'].mean().unstack('platform')
segment_trends.plot(figsize=(10,5), title='Engagement Rate by Platform and Month', marker='o')
plt.ylabel('Engagement Rate (%)')
plt.xlabel('Month')
plt.tight_layout()
plt.show()

# Use 14-day rolling average for engagement rate
time_series_df['rolling_engagement'] = time_series_df['engagement_rate'].rolling(window=14, min_periods=1).mean()
plt.plot(time_series_df['date'], time_series_df['engagement_rate'], color='lightgray', label='Daily Engagement Rate')
plt.plot(time_series_df['date'], time_series_df['rolling_engagement'], color='red', label='14-day Rolling Avg')
plt.title('Engagement Rate Trend with Rolling Average')
plt.xlabel('Date')
plt.ylabel('Engagement Rate (%)')
plt.legend()
plt.tight_layout()
plt.show()

# Randomly introduce missing values for demonstration
sm_content_df.loc[sm_content_df.sample(frac=0.01).index, 'likes'] = np.nan
missing_counts = sm_content_df.isnull().sum()
print('Number of missing values in each column:')
print(missing_counts)

# Fill or drop missing values to clean data for later calculations
sm_content_df['likes'] = sm_content_df['likes'].fillna(0)
print(sm_content_df['likes'].isnull().sum())

# Bad practice: Aggregate engagement without grouping
bad_sum = sm_content_df['likes'].sum()
print('Sum of all likes (across all posts):', bad_sum)

# Correct way: Group by platform before aggregating
correct_sum = sm_content_df.groupby('platform')['likes'].sum()
print('Total likes by platform:')
print(correct_sum)

# Consistency check: Engagement rate when views are zero
test_df = sm_content_df.copy()
test_df.loc[0, 'views'] = 0   # Set up a zero-views example
try:
    test_df['test_engagement'] = (test_df['likes'] + test_df['comments'] + test_df['shares']) / test_df['views'] * 100
    print(test_df.loc[0, 'test_engagement'])
except ZeroDivisionError:
    print('Error: Division by zero occurred!')

# Safe division: Replace zero views with np.nan before division
test_df['safe_engagement'] = np.where(test_df['views'] == 0, np.nan,
    (test_df['likes'] + test_df['comments'] + test_df['shares']) / test_df['views'] * 100)
print(test_df[['views', 'safe_engagement']].head(3))

# Tiny End-to-End Problem: Recommend a Content Strategy
# Step 1: Identify platform with highest median engagement rate
platform_median = sm_content_df.groupby('platform')['engagement_rate'].median().sort_values(ascending=False)
best_platform = platform_median.index[0]
print('Highest-median platform:', best_platform)

# Step 2: Find the best posting month for that platform
best_month = sm_content_df[sm_content_df['platform'] == best_platform].groupby('month')['engagement_rate'].mean().idxmax()
print('Best month to post on', best_platform, 'is', best_month)

# Step 3: Find sample posts to study for future inspiration
sample_posts = sm_content_df[(sm_content_df['platform'] == best_platform) & (sm_content_df['month'] == best_month)].sort_values('engagement_rate', ascending=False).head(3)
print('Top example posts:')
print(sample_posts[['post_id', 'views', 'engagement_rate']])