1105_4.10 製作一個 MP4

import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
from matplotlib.dates import DateFormatter, date2num, num2date
import numpy as np

# Read the CSV file
input_file = '/content/drive/My Drive/TKU_Industry_Analysis_Semiconductor/Dataset_SEMI/semiconductor_merged_1016.csv'
df = pd.read_csv(input_file)

# Define company abbreviations
company_abbr = {
    'ASML Holding N.V.': 'ASML',
    'Advanced Micro Devices, Inc.': 'AMD',
    'Analog Devices, Inc.': 'ADI',
    'Applied Materials, Inc.': 'AMAT',
    'Arm Holdings plc': 'ARM',
    'Broadcom Inc.': 'AVGO',
    'Intel Corporation': 'INTC',
    'KLA Corporation': 'KLA',
    'Lam Research Corporation': 'LRCX',
    'Marvell Technology, Inc.': 'MRVL',
    'MediaTek Inc.': 'MTK',
    'Micron Technology, Inc.': 'MU',
    'NVIDIA Corporation': 'NVDA',
    'QUALCOMM Incorporated': 'QCOM',
    'SK hynix Inc.': 'SKH',
    'Samsung Electronics Co., Ltd.': 'SEC',
    'Synopsys, Inc.': 'SNPS',
    'Taiwan Semiconductor Manufacturing Company Limited': 'TSMC',
    'Texas Instruments Incorporated': 'TXN',
    'Tokyo Electron Limited': 'TEL'
}

# Replace company names with abbreviations
df['Company'] = df['Company Name'].replace(company_abbr)

# Ensure 'Date' column is in datetime format
df['Date'] = pd.to_datetime(df['Date'])

# Filter data from January 1, 2024
df = df[df['Date'] >= '2024-01-01']

# Calculate daily closing price percentage change
df.loc[:, 'Close_Price_Change%'] = df.groupby('Company Name')['Close Price_Local'].pct_change() * 100

# Filter required columns
df_filtered = df[['Date', 'Company', 'Close_Price_Change%']].dropna()

# Define highlighted companies with their colors and line styles
highlighted_companies = {
    'NVDA': {'color': 'red', 'linewidth': 2},
    'TSMC': {'color': 'blue', 'linewidth': 2},
    'AMAT': {'color': 'lime', 'linewidth': 2},
    'ASML': {'color': 'darkorchid', 'linewidth': 2},
    'SEC': {'color': 'lightseagreen', 'linewidth': 2},
}

# Create the animation plot
fig, ax = plt.subplots(figsize=(10, 6))

# Set y-axis limits
ax.set_ylim(-15, 15)
ax.set_xlabel('Date')
ax.set_ylabel('Close Price Change %')

# Set date formatting
ax.xaxis.set_major_formatter(DateFormatter('%Y-%m-%d'))

# Rotate x-axis labels
plt.xticks(rotation=90)

# Initialize lines for each company
lines = {}
for company in df_filtered['Company'].unique():
    if company in highlighted_companies:
        # Set specific company colors and line styles
        lines[company] = ax.plot([], [], label=company, color=highlighted_companies[company]['color'],
                                 linewidth=highlighted_companies[company]['linewidth'])[0]
    else:
        # Set remaining companies with low opacity
        lines[company] = ax.plot([], [], label=company, color='gray', alpha=0.1, linewidth=1)[0]

# Set legend
ax.legend()

# Update function for each frame
def update(frame):
    current_date = df_filtered['Date'].unique()[frame]
    filtered_data = df_filtered[df_filtered['Date'] <= current_date]

    for company in lines.keys():
        company_data = filtered_data[filtered_data['Company'] == company]
        lines[company].set_data(company_data['Date'], company_data['Close_Price_Change%'])

    ax.set_title(f"Date: {current_date.strftime('%Y-%m-%d')}")
    ax.relim()  # Recalculate limits
    ax.autoscale_view()  # Update view limits

    # Update regression lines for NVDA and AMAT
    for company, color, linestyle in zip(['NVDA', 'AMAT'], ['black', 'gray'], ['-', '--']):
        company_data = df_filtered[df_filtered['Company'] == company]
        if not company_data.empty:
            x = date2num(company_data['Date'])
            y = company_data['Close_Price_Change%']
            coefficients = np.polyfit(x, y, 2)  # Fit a second-degree polynomial
            polynomial = np.poly1d(coefficients)

            # Generate points for the regression line
            x_reg = np.linspace(x.min(), x.max(), 100)
            y_reg = polynomial(x_reg)

            # Plot the regression line
            ax.plot(num2date(x_reg), y_reg, color=color, linestyle=linestyle, linewidth=2)

# Create the animation
ani = FuncAnimation(fig, update, frames=len(df_filtered['Date'].unique()), repeat=False, interval=100)

# Save as MP4
output_file = '/content/drive/My Drive/TKU_Industry_Analysis_Semiconductor/Dataset_SEMI/testing_3.mp4'
ani.save(output_file, writer='ffmpeg', fps=10)

plt.close()