google_trends_analysis.py

import os
import pandas as pd
import matplotlib.pyplot as plt
from statsmodels.tsa.seasonal import STL

# Create a directory for analysis output
output_dir = 'analysis_output'
if not os.path.exists(output_dir):
    os.makedirs(output_dir)

# Load combined data
combined_data = pd.read_csv('combined_trends_data.csv', index_col=0, parse_dates=True)

# Define key global events related to climate change
global_events = {
    pd.Timestamp('2009-12-01'): 'COP15 Copenhagen',
    pd.Timestamp('2015-12-01'): 'Paris Agreement',
    pd.Timestamp('2021-11-01'): 'COP26 Glasgow'
}

# Function to perform STL decomposition and plot
def analyze_city_data(city_data, city_name):
    for keyword in city_data.columns:
        if keyword == 'City':
            continue
        
        # Perform STL decomposition
        stl = STL(city_data[keyword], seasonal=13)
        result = stl.fit()
        
        # Plot the results
        fig, ax = plt.subplots(figsize=(10, 6))
        ax.plot(city_data.index, result.trend, label='Trend', color='blue')
        ax.set_title(f'STL Decomposition of {keyword} in {city_name}')
        ax.set_xlabel('Year')
        ax.set_ylabel('Trend')
        
        # Annotate global events
        for date, event in global_events.items():
            if date in result.trend.index:
                ax.annotate(event, xy=(date, result.trend.loc[date]), xytext=(date, result.trend.loc[date] + 5),
                            arrowprops=dict(facecolor='black', shrink=0.05), fontsize=8, rotation=45)
        
        plt.legend()
        plt.tight_layout()
        
        # Save the plot
        plt.savefig(f'{output_dir}/{city_name.lower().replace(" ", "_")}_{keyword}_trend.png')
        plt.close()

# Analyze data for each city
for city in combined_data['City'].unique():
    city_data = combined_data[combined_data['City'] == city].drop(columns=['City'])
    analyze_city_data(city_data, city)

print("Analysis complete. Check the 'analysis_output' directory for results.")