Tesla Time Series Analysis

# Import necessary libraries
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from statsmodels.tsa.stattools import adfuller

# Load the data
engagement_data = pd.read_csv('eps_data.csv', parse_dates=['date'], index_col='date')

# convert in the index Date into the DateTime format
engagement_data.index = pd.DatetimeIndex(engagement_data.index.values,freq=engagement_data.index.inferred_freq)

#Sort DataFrame 
engagement_data = engagement_data.sort_index()

# Check for missing values and handle if necessary
engagement_data.isnull().sum()

# Perform stationarity tests and difference the data if necessary
def adf_test(series):
    result = adfuller(series)
    print('ADF Statistic:', result[0])
    print('p-value:', result[1])
    print('Critical Values:')
    for key, value in result[4].items():
        print(f'\t{key}: {value}')
        
adf_test(engagement_data['EPS'])

# If non-stationary, difference the data
engagement_data['EPS_diff'] = engagement_data['EPS'].diff().dropna()

# Drop the NaN values after differencing
engagement_data.dropna(inplace=True)

# Correlation analysis
correlation_matrix = engagement_data.corr()
print(correlation_matrix)

# Visualization
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm')
plt.title('Correlation Heatmap')
plt.show()

sns.pairplot(engagement_data)
plt.title('Pair Plot')
plt.show()

# Import necessary libraries
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression  
from sklearn.metrics import mean_squared_error, r2_score

# Load the data
engagement_data = pd.read_csv('eps_data.csv', parse_dates=['date'], index_col='date')

# convert in the index Date into the DateTime format
engagement_data.index = pd.DatetimeIndex(engagement_data.index.values,freq=engagement_data.index.inferred_freq)

#Sort DataFrame 
engagement_data = engagement_data.sort_index()

# Check for missing values and handle if necessary
engagement_data.isnull().sum()

# Drop the NaN values after differencing
engagement_data.dropna(inplace=True)

# Search Trend Analysis 

# Extract EPS and the engagement metric
X = engagement_data[['Google Search Trends']]  # Input feature (Google Search Trends) - Modified to be a DataFrame
y = engagement_data['EPS']  # Target variable (EPS)

# Split the data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Initialize and fit the linear regression model
model = LinearRegression()  # Corrected model type to LinearRegression
model.fit(X_train, y_train)

# Make predictions on the testing set
y_pred = model.predict(X_test)

# Calculate evaluation metrics
mse = mean_squared_error(y_test, y_pred)
r2 = r2_score(y_test, y_pred)

print("Mean Squared Error:", mse)
print("R-squared:", r2)

# Plot the regression line
plt.figure(figsize=(8, 6))
plt.scatter(X_test, y_test, color='blue', label='Actual')
plt.plot(X_test, y_pred, color='red', label='Predicted')
plt.title('Linear Regression: EPS vs. Google Search Trends')
plt.xlabel('Google Search Trends')
plt.ylabel('EPS')
plt.legend()
plt.show()

# Import necessary libraries
import pandas as pd
from statsmodels.tsa.vector_ar.var_model import VAR
from statsmodels.tsa.stattools import adfuller
import matplotlib.pyplot as plt
from math import sqrt
from sklearn.metrics import mean_squared_error

# Load the data
engagement_data = pd.read_csv('eps_data.csv', parse_dates=['date'], index_col='date')

# convert in the index Date into the DateTime format
engagement_data.index = pd.DatetimeIndex(engagement_data.index.values,freq=engagement_data.index.inferred_freq)

#Sort DataFrame 
engagement_data = engagement_data.sort_index()

# Check for missing values and handle if necessary
engagement_data.isnull().sum()

# Perform stationarity tests and difference the data if necessary
def adf_test(series):
    result = adfuller(series)
    print('ADF Statistic:', result[0])
    print('p-value:', result[1])
    print('Critical Values:')
    for key, value in result[4].items():
        print(f'\t{key}: {value}')
        
adf_test(engagement_data['EPS'])

# If non-stationary, difference the data
engagement_data['EPS_diff'] = engagement_data['EPS'].diff().dropna()

# Drop the NaN values after differencing
engagement_data.dropna(inplace=True)


# Fit VAR model on the differenced data. I've chosen this model because VAR offers modelling multiple time series variables simultaneously, you can still use a VAR model with one dependent variable and one independent variable. This approach allows you to capture any potential feedback effects between EPS and the stakeholder engagement metrics.

model = VAR(train_data[['EPS_diff', 'Twitter Sentiment','Google News Trends','Google Search Trends']])
results = model.fit(3)

# Forecast
forecast_steps = len(test_data)
forecast = results.forecast(y=results.endog, steps=forecast_steps)

# Invert differencing if necessary
# If differenced data was used, invert the differencing
def invert_diff(original, diff_series):
    return original + diff_series.cumsum()

forecast = pd.DataFrame(forecast, index=test_data.index, columns=['EPS_diff_forecast', 'Twitter_forecast','News_forecast','Search_forecast'])
forecast['EPS_forecast'] = invert_diff(train_data['EPS'].iloc[-1], forecast['EPS_diff_forecast'])

# Evaluate model
mse = mean_squared_error(test_data['EPS'], forecast['EPS_forecast'])
rmse = sqrt(mse)
print('RMSE:', rmse)

# Plot actual vs forecasted EPS
plt.figure(figsize=(12, 6))
plt.plot(test_data.index, test_data['EPS'], label='Actual EPS', color='blue')
plt.plot(forecast.index, forecast['EPS_forecast'], label='Forecasted EPS', color='red')
plt.title('Actual vs Forecasted EPS')
plt.xlabel('Date')
plt.ylabel('EPS')
plt.legend()
plt.show()