Navigate below Welcome sidebar:
", unsafe_allow_html=True) st.sidebar.markdown("[Welcome](#welcome-section)") # Welcome section st.markdown( """The FIFA World Cup is the biggest football sports competition where countries from all over the world come together to compete for the most glorious and amazing cup. π In this app, we're diving into what affects how many goals a team scores in every game during the FIFA World Cup 2022, & Other factors which matters in The Football Match. Why? Well, in football, by scoring more goals often means you're more likely to win the game. Let's explore why that's the case.
"Football is about scoring goals." - Pep Guardiola
Navigate below Introduction sidebar:
", unsafe_allow_html=True) # Welcoming message and image st.markdown("π‘ Pro Tip:
", unsafe_allow_html=True) st.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you, in the left side bar! π Feel free to adjust the volume π or stop the song βΉοΈ whenever you want. πΊπ
", unsafe_allow_html=True) st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.audio(audio_fifa_3, format='audio/mp3') st.video(video_intro, format='video/mp4') # Objectives st.header("π― Objectives") st.markdown(""" Our goal is to analyze key factors affecting team performance in the FIFA World Cup 2022. We're particularly interested in understanding what influences the number of goals scored by each team. We're also exploring other factors like possession to gain insights into team dynamics and strategies. """) # Key Variables st.markdown("### Key Variables") st.markdown("Below are the key variables we emphasize in our analysis, though there are more additional variables considered:") st.markdown("- Team") st.markdown("- Possession") st.markdown("- Number of Goals") st.markdown("- Corners") st.markdown("- On Target Attempts") st.markdown("- Defensive Pressures Applied") # Description of Data st.markdown("### Description of Data") st.markdown("Let's take a look at some descriptive statistics of the data:") # Load data df = pd.read_csv("FIFAWorldCup2022.csv") # Interactive widgets st.sidebar.title('Data Exploration Options') # Default selection for Team 1 default_selected_team1 = ['QATAR'] # Dropdown menu for team selection (Team 1) selected_teams_team1 = st.sidebar.multiselect('Select Teams (Team 1)', df['team1'].unique(), default=default_selected_team1) # Default selection for Team 2 default_selected_team2 = ['ECUADOR'] # Dropdown menu for team selection (Team 2) selected_teams_team2 = st.sidebar.multiselect('Select Teams (Team 2)', df['team2'].unique(), default=default_selected_team2) # Filter data based on user selections for both teams filtered_df_team1 = df[df['team1'].isin(selected_teams_team1)] filtered_df_team2 = df[df['team2'].isin(selected_teams_team2)] # Combine filtered data for both teams filtered_df = pd.concat([filtered_df_team1, filtered_df_team2]) # Display interactive report if st.sidebar.button('Show Report'): if filtered_df.empty: st.warning("No data available for the selected teams.") else: # Summary statistics st.subheader("Summary Statistics") st.write(filtered_df.describe()) # Bar chart for number of goals st.subheader("Number of Goals Comparison") fig_goals = px.bar(filtered_df, x='team1', y='number of goals team1', color='team1', title='Number of Goals Comparison') st.plotly_chart(fig_goals) # Histogram for possession st.subheader("Distribution of Possession") fig_possession = px.histogram(filtered_df, x='possession team1', color='team1', nbins=20, title='Possession Distribution') st.plotly_chart(fig_possession) # Line plot for trends over time (assuming 'date' column represents time) if 'date' in filtered_df.columns: st.subheader("Trends Over Time") fig_trends = px.line(filtered_df, x='date', y='possession team1', color='team1', title='Possession Over Time') st.plotly_chart(fig_trends) # Additional statistics and insights st.subheader("Additional Statistics and Insights") # Remove percentage symbols and convert to numeric filtered_df['possession team1'] = filtered_df['possession team1'].str.replace('%', '').astype(float) # Create a bar plot for total goals scored, average possession, and average number of goals per game fig, ax = plt.subplots(figsize=(10, 6)) # Total goals scored total_goals = filtered_df['number of goals team1'].sum() ax.bar("Total Goals Scored", total_goals, color='blue') ax.text("Total Goals Scored", total_goals, f'{total_goals}', ha='center', va='bottom') # Average possession avg_possession = filtered_df['possession team1'].mean() ax.bar("Average Possession", avg_possession, color='green') ax.text("Average Possession", avg_possession, f'{avg_possession:.2f}%', ha='center', va='bottom') # Average number of goals per game avg_goals_per_game = filtered_df['number of goals team1'].mean() ax.bar("Average Goals Per Game", avg_goals_per_game, color='orange') ax.text("Average Goals Per Game", avg_goals_per_game, f'{avg_goals_per_game:.2f}', ha='center', va='bottom') # Set labels and title ax.set_ylabel('Value') ax.set_title('Comparison of Statistics') plt.xticks(rotation=45) plt.tight_layout() # Display the plot st.pyplot(fig) # Dynamic data exploration for team 1 st.subheader("Dynamic Data Exploration (Team 1)") st.write(filtered_df_team1) # Dynamic data exploration for team 2 st.subheader("Dynamic Data Exploration (Team 2)") st.write(filtered_df_team2) # User Feedback Integration st.sidebar.title('User Feedback') user_email = st.sidebar.text_input("Enter your email address:") feedback = st.sidebar.text_area("Please provide your feedback here:") submit_button = st.sidebar.button("Submit Feedback") if submit_button: # Store feedback in a file or database with open("feedback.txt", "a") as f: f.write("Email: {}\nFeedback: {}\n".format(user_email, feedback)) st.sidebar.success("Thank you for your feedback!") # Send feedback to email sender_email = user_email # Use user's email as sender receiver_emails = ["jackson.mukeshimana@nyu.edu", "mukesjackson02@gmail.com"] # Update with receiver emails # Compose email message = MIMEMultipart() message["From"] = sender_email message["To"] = ", ".join(receiver_emails) message["Subject"] = "User Feedback" # Add message body message.attach(MIMEText("User Email: {}\n\nFeedback: {}".format(user_email, feedback), "plain")) # Connect to SMTP server and send email with smtplib.SMTP("smtp.gmail.com", 587) as server: server.starttls() server.sendmail(sender_email, receiver_emails, message.as_string()) st.sidebar.success("Your feedback has been submitted and sent to the admins.") # Convert categorical columns to numeric codes df['team1'] = df['team1'].astype('category').cat.codes df['team2'] = df['team2'].astype('category').cat.codes # Remove percentage signs and convert to numeric columns_to_convert = ['possession team1', 'possession team2', 'possession in contest'] for column in columns_to_convert: df[column] = df[column].astype(str).str.rstrip('%').astype(float) # Convert converted columns to categorical codes for column in columns_to_convert: df[column] = df[column].astype('category').cat.codes # Convert other categorical columns to numeric codes columns_to_convert_to_codes = ['date', 'hour', 'category'] for column in columns_to_convert_to_codes: df[column] = df[column].astype('category').cat.codes # Display summary statistics st.dataframe(df.describe()) # Convert 'date' column to datetime df['date'] = pd.to_datetime(df['date']) # Missing Values st.markdown("### Missing Values") st.markdown("Let's examine the presence of missing values in our dataset:") # Calculate percentage of missing values for each column missing_values = df.isnull().sum() / len(df) * 100 # Display missing value percentages st.write("Percentage of missing values for each column:") st.write(missing_values) # Assess overall completeness of the dataset completeness_ratio = df.notnull().sum().sum() / (len(df) * len(df.columns)) st.write(f"Overall completeness ratio: {completeness_ratio:.2f}") if completeness_ratio >= 0.85: st.success("The dataset has a high level of completeness, providing us with reliable data for analysis.") else: st.warning("The dataset has a low level of completeness, which may affect the reliability of our analysis.") # Conclusion st.markdown("### Recap") st.markdown("In this dashboard page, we explored the FIFA World Cup 2022 dataset. We've seen the key variables like possession, number of goals team1, corners, defensive pressures applied, and more others. We checked also the Cleanliness of our data set and checked any missing values maybe in our data set, for reliability and usability purposes.") # Visualization Page elif st.session_state.app_mode == 'Visualization': # Play FIFA World Cup song st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you! πΆ Feel free to adjust the volume or stop the song whenever you want. π§
", unsafe_allow_html=True) st.sidebar.audio(audio_fifa_2, format='audio/mp3') # Left sidebar for text st.subheader("Explore visualizations of the FIFA World Cup 2022 data.") st.image(image_quatar2022, width=800) # Load the FIFA World Cup 2022 dataset df = pd.read_csv('FIFAWorldCup2022.csv') # Convert categorical columns to numeric codes df['team1'] = df['team1'].astype('category').cat.codes df['team2'] = df['team2'].astype('category').cat.codes # Remove percentage signs and convert to numeric columns_to_convert = ['possession team1', 'possession team2', 'defensive pressures applied team1', 'passes team2', 'passes completed team2', 'on target attempts team2', 'inbehind offers to receive team2', 'attempted defensive line breaks team2'] for column in columns_to_convert: df[column] = df[column].astype(str).str.rstrip('%').astype(float) # Convert converted columns to categorical codes for column in columns_to_convert: df[column] = df[column].astype('category').cat.codes # Convert other categorical columns to numeric codes columns_to_convert_to_codes = ['date', 'hour', 'category'] for column in columns_to_convert_to_codes: df[column] = df[column].astype('category').cat.codes # Select independent variables and target variable selected_features = ["assists team2", "attempted defensive line breaks team2", "on target attempts team2", "inbehind offers to receive team2", "possession team2", "passes completed team2", "number of goals team2"] # Extract selected columns from the dataset df_selected = df[selected_features] # Calculate the correlation matrix corr_matrix = df_selected.corr() # Plot the heatmap using Streamlit st.write("## Correlation Heatmap") st.markdown("{i}. {fact}
", unsafe_allow_html=True) # Conclusion and Surprise Element show_conclusion = st.button("Show Conclusion and Surprise Element") if show_conclusion: st.subheader("Conclusion") st.write("Congratulations! You've explored a variety of visualizations and interactive features to gain insights from the FIFA World Cup 2022 dataset. But wait, there's more!") # Surprise Element: Random Fun Fact random_fact = "Did you know that the FIFA World Cup trophy weighs about 6.175 kilograms (13.61 pounds)?" st.markdown(f"Here's a random fun fact: {random_fact}
", unsafe_allow_html=True) elif st.session_state.app_mode == 'Prediction': st.subheader("Prediction") st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you! πΆ Feel free to adjust the volume or stop the song whenever you want. π§
", unsafe_allow_html=True) st.sidebar.audio(audio_fifa_4, format='audio/mp3') st.image(image_featured, use_column_width=True) st.title("FIFA World Cup 2022 Data Analysis - Prediction") st.markdown("Select a machine learning model and variables to predict outcomes.") # Load the dataset df = pd.read_csv('FIFAWorldCup2022.csv') # Convert categorical columns to numeric codes df['team1'] = df['team1'].astype('category').cat.codes df['team2'] = df['team2'].astype('category').cat.codes # Remove percentage signs and convert to numeric columns_to_convert = ['possession team1', 'possession team2', 'possession in contest'] for column in columns_to_convert: df[column] = df[column].astype(str).str.rstrip('%').astype(float) # Convert converted columns to categorical codes for column in columns_to_convert: df[column] = df[column].astype('category').cat.codes # Convert other categorical columns to numeric codes columns_to_convert_to_codes = ['date', 'hour', 'category'] for column in columns_to_convert_to_codes: df[column] = df[column].astype('category').cat.codes # Set dependent variable selected_target = 'number of goals team2' # Set default independent variables default_independent_variables = ["assists team2", "attempts inside the penalty area team2", "offsides team2"] # Calculate correlation with the dependent variable corr_with_target = df.corr()[selected_target].abs() # Filter out independent variables with correlation > 0.1 filtered_features = corr_with_target[corr_with_target > 0.1].index.tolist() # Ensure default values exist in available options default_features = [feat for feat in default_independent_variables if feat in filtered_features] # Features and target variable selection selected_features = st.multiselect("Select Independent Variables", filtered_features, default=default_features) # Machine learning model selection selected_models = st.multiselect("Select Model(s)", ['Linear Regression', 'Random Forest', 'Gradient Boosting'], default=['Linear Regression']) # Custom hyperparameters for selected models custom_hyperparameters = {} for model in selected_models: if model == 'Random Forest': custom_hyperparameters['Random Forest'] = { 'n_estimators': st.number_input("Number of Estimators (Random Forest)", min_value=10, max_value=1000, value=100, step=10) } elif model == 'Gradient Boosting': custom_hyperparameters['Gradient Boosting'] = { 'n_estimators': st.number_input("Number of Estimators (Gradient Boosting)", min_value=10, max_value=1000, value=100, step=10), 'learning_rate': st.number_input("Learning Rate (Gradient Boosting)", min_value=0.01, max_value=1.0, value=0.1, step=0.01) } if not selected_features: st.warning("Please select at least one independent variable.") else: # Extract selected columns from the dataset df_selected = df[selected_features + [selected_target]] # Remove rows with missing values df_selected = df_selected.dropna() if df_selected.empty: st.warning("No data available after removing rows with missing values. Please choose different variables.") else: # Check if selected variables have numeric data numeric_columns = df_selected.select_dtypes(include=['float', 'int']).columns if len(numeric_columns) != len(selected_features) + 1: # Check if all selected variables are numeric non_numeric_variables = [var for var in selected_features + [selected_target] if var not in numeric_columns] st.error(f"The following selected variables contain non-numeric values: {', '.join(non_numeric_variables)}") else: X = df_selected[selected_features] y = df_selected[selected_target] # Split 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) # Guiding message st.info("Select a machine learning model.") # Display dependent variable with enhanced style st.markdown(f"Dependent Variable to Predict: {selected_target}
", unsafe_allow_html=True) for model in selected_models: st.subheader(f"{model} Model") if model == 'Linear Regression': # Linear Regression model implementation try: # Train the model model = LinearRegression() model.fit(X_train, y_train) # Make predictions y_pred = model.predict(X_test) # Evaluate the model r2 = r2_score(y_test, y_pred) mse = mean_squared_error(y_test, y_pred) rmse = np.sqrt(mse) # Model Performance Visualization st.subheader("Model Performance Visualization") # Create histogram data hist_data = [y_test, y_pred] group_labels = ['Actual', 'Predicted'] # Create the histogram using Plotly fig_pred_actual_hist = ff.create_distplot(hist_data, group_labels, bin_size=0.5, colors=['blue', 'orange']) # Update layout with enhanced features fig_pred_actual_hist.update_layout( title='Predicted vs Actual Histogram', xaxis_title='Values', yaxis_title='Frequency', showlegend=True, plot_bgcolor='rgba(255, 255, 255, 0.9)', # Set background color template='plotly_white', # Use white template for better contrast width=800, # Increase default width of the chart height=600, # Increase default height of the chart ) # Add buttons for interactivity fig_pred_actual_hist.update_layout( updatemenus=[ { 'buttons': [ { 'args': [None, {'frame': {'duration': 500, 'redraw': True}, 'fromcurrent': True}], 'label': 'Play', 'method': 'animate' }, { 'args': [[None], {'frame': {'duration': 0, 'redraw': True}, 'mode': 'immediate', 'transition': {'duration': 0}}], 'label': 'Pause', 'method': 'animate' } ], 'direction': 'left', 'pad': {'r': 10, 't': 10}, 'showactive': False, 'type': 'buttons', 'x': 0.05, # Adjust position of the buttons 'xanchor': 'right', 'y': 1.1, # Adjust position of the buttons 'yanchor': 'top' }, { 'buttons': [ {'args': [None, {'xaxis': {'type': 'linear'}, 'yaxis': {'type': 'linear'}}], 'label': 'Reset Zoom', 'method': 'relayout'} ], 'direction': 'down', 'showactive': False, 'type': 'buttons', 'x': 0.05, # Adjust position of the buttons 'xanchor': 'right', 'y': 1.05, # Adjust position of the buttons 'yanchor': 'top' } ] ) # Display the histogram st.plotly_chart(fig_pred_actual_hist) # Display model performance metrics st.subheader("Model Performance Metrics") st.write(f"{model} Model Performance:") st.write(f"R-squared: {r2:.2f}") st.write(f"Mean Squared Error: {mse:.2f}") st.write(f"Root Mean Squared Error: {rmse:.2f}") st.write("Interpretation:") if r2 >= 0.7: st.info(f"R-squared of {r2:.2f} shows that the model explains a large proportion of the variance in the dependent variable, indicating a strong relationship between the selected features and the number of goals of the team.") elif r2 >= 0.5: st.warning(f"R-squared of {r2:.2f} shows that the model explains a moderate proportion of the variance in the dependent variable, suggesting a moderate relationship between the selected features and the number of goals of the team.") else: st.error(f"R-squared of {r2:.2f} shows that the model does not explain much of the variance in the dependent variable, indicating a weak relationship between the selected features and the number of goals of the team.") # Check if R-squared is less than zero if r2 < 0: st.error("R-squared is less than zero. There may be an issue with the chosen variable in the dataset. Please consider removing this variable.") except ValueError as e: st.error(f"Error: {e}. Please ensure all selected variables are numeric.") elif model == 'Random Forest': # Random Forest model implementation with custom hyperparameters try: # Train the model with custom hyperparameters n_estimators = custom_hyperparameters['Random Forest']['n_estimators'] model = RandomForestRegressor(n_estimators=n_estimators) model.fit(X_train, y_train) # Make predictions y_pred = model.predict(X_test) # Evaluate the model r2 = r2_score(y_test, y_pred) mse = mean_squared_error(y_test, y_pred) rmse = np.sqrt(mse) # Create scatter plot data scatter_data = go.Scatter(x=y_test, y=y_pred, mode='markers', name='Predicted vs Actual', marker=dict(color='orange')) # Create perfect prediction line data perfect_line = go.Scatter(x=[y_test.min(), y_test.max()], y=[y_test.min(), y_test.max()], mode='lines', name='Perfect Prediction', line=dict(color='blue', dash='dash')) # Create the figure fig_rf = go.Figure(data=[scatter_data, perfect_line]) # Update layout with enhanced features fig_rf.update_layout( title='Random Forest: Predicted vs Actual', xaxis_title='Actual', yaxis_title='Predicted', showlegend=True, plot_bgcolor='rgba(255, 255, 255, 0.9)', # Set background color xaxis=dict(showgrid=True, gridcolor='lightgray'), # Show gridlines on x-axis yaxis=dict(showgrid=True, gridcolor='lightgray'), # Show gridlines on y-axis hovermode='closest', # Set hover mode to show closest data point template='plotly_white', # Use white template for better contrast width=900, # Increase default width of the chart height=700, # Increase default height of the chart ) # Add buttons for interactivity fig_rf.update_layout( updatemenus=[ { 'buttons': [ {'args': [None, {'frame': {'duration': 500, 'redraw': True}, 'fromcurrent': True}], 'label': 'Play', 'method': 'animate'}, {'args': [[None], {'frame': {'duration': 0, 'redraw': True}, 'mode': 'immediate', 'transition': {'duration': 0}}], 'label': 'Pause', 'method': 'animate'} ], 'direction': 'left', 'pad': {'r': 10, 't': 10}, 'showactive': False, 'type': 'buttons', 'x': 0.05, # Adjust position of the buttons 'xanchor': 'right', 'y': 1.1, # Adjust position of the buttons 'yanchor': 'top' }, { 'buttons': [ {'args': [None, {'xaxis': {'type': 'linear'}, 'yaxis': {'type': 'linear'}}], 'label': 'Reset Zoom', 'method': 'relayout'} ], 'direction': 'down', 'showactive': False, 'type': 'buttons', 'x': 0.05, # Adjust position of the buttons 'xanchor': 'right', 'y': 1.05, # Adjust position of the buttons 'yanchor': 'top' } ] ) # Display the scatter plot st.plotly_chart(fig_rf) # Display model performance metrics st.subheader("Model Performance Metrics") st.write(f"{model} Model Performance:") st.write(f"R-squared: {r2:.2f}") st.write(f"Mean Squared Error: {mse:.2f}") st.write(f"Root Mean Squared Error: {rmse:.2f}") st.write("Interpretation:") if r2 >= 0.7: st.info(f"R-squared of {r2:.2f} shows that the model explains a large proportion of the variance in the dependent variable, indicating a strong relationship between the selected features and the number of goals of the team.") elif r2 >= 0.5: st.warning(f"R-squared of {r2:.2f} shows that the model explains a moderate proportion of the variance in the dependent variable, suggesting a moderate relationship between the selected features and the number of goals of the team.") else: st.error(f"R-squared of {r2:.2f} shows that the model does not explain much of the variance in the dependent variable, indicating a weak relationship between the selected features and the number of goals of the team.") # Check if R-squared is less than zero if r2 < 0: st.error("R-squared is less than zero. There may be an issue with the chosen variable in the dataset. Please consider removing this variable.") except ValueError as e: st.error(f"Error: {e}. Please ensure all selected variables are numeric.") elif model == 'Gradient Boosting': # Gradient Boosting model implementation with custom hyperparameters try: # Train the model with custom hyperparameters n_estimators = custom_hyperparameters['Gradient Boosting']['n_estimators'] learning_rate = custom_hyperparameters['Gradient Boosting']['learning_rate'] model = GradientBoostingRegressor(n_estimators=n_estimators, learning_rate=learning_rate) model.fit(X_train, y_train) # Make predictions y_pred = model.predict(X_test) # Evaluate the model r2 = r2_score(y_test, y_pred) mse = mean_squared_error(y_test, y_pred) rmse = np.sqrt(mse) # Create a 3D scatter plot using Plotly fig = go.Figure(data=[go.Scatter3d( x=y_test, y=y_pred, z=X_test['assists team2'], # Use a feature as the third dimension for added insight mode='markers', marker=dict( size=5, color=X_test['assists team2'], # Color points by a feature colorscale='Viridis', # Choose a color scale opacity=0.8 ) )]) # Update layout for better presentation fig.update_layout( scene=dict( xaxis_title='Actual Goals of Team', yaxis_title='Predicted Goals of Team', zaxis_title='Assists of Team', # Update axis titles bgcolor='rgba(139, 69, 19, 0.8)', # Set brown background color of the 3D scene ), title=dict(text='Gradient Boosting: Actual vs Predicted Goals of Teams', x=0.5), # Update title margin=dict(l=0, r=0, b=0, t=0), # Update margin for better layout width=900, # Increase default width of the chart height=700, # Increase default height of the chart ) # Display the 3D scatter plot st.plotly_chart(fig) # Display model performance metrics st.subheader("Model Performance Metrics") st.write(f"{model} Model Performance:") st.write(f"R-squared: {r2:.2f}") st.write(f"Mean Squared Error: {mse:.2f}") st.write(f"Root Mean Squared Error: {rmse:.2f}") st.write("Interpretation:") if r2 >= 0.7: st.info(f"R-squared of {r2:.2f} shows that the model explains a large proportion of the variance in the dependent variable, indicating a strong relationship between the selected features and the number of goals of the team.") elif r2 >= 0.5: st.warning(f"R-squared of {r2:.2f} shows that the model explains a moderate proportion of the variance in the dependent variable, suggesting a moderate relationship between the selected features and the number of goals of the team.") else: st.error(f"R-squared of {r2:.2f} shows that the model does not explain much of the variance in the dependent variable, indicating a weak relationship between the selected features and the number of goals of the team.") # Check if R-squared is less than zero if r2 < 0: st.error("R-squared is less than zero. There may be an issue with the chosen variable in the dataset. Please consider removing this variable.") except ValueError as e: st.error(f"Error: {e}. Please ensure all selected variables are numeric.") elif st.session_state.app_mode == 'Feature of Importance & Shap': st.subheader("Features of Importance & Shap") st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you! πΆ Feel free to adjust the volume or stop the song whenever you want. π§
", unsafe_allow_html=True) st.sidebar.audio(audio_fifa_5, format='audio/mp3') st.image(image_F, width=800) st.title("Feature of Importance & Shap") df = pd.read_csv('FIFAWorldCup2022.csv') # Preprocess the data df['team1'] = df['team1'].astype('category').cat.codes df['team2'] = df['team2'].astype('category').cat.codes columns_to_convert = ['possession team1', 'possession team2', 'possession in contest'] for column in columns_to_convert: df[column] = df[column].str.rstrip('%').astype(float).astype('category').cat.codes columns_to_convert_to_codes = ['date', 'hour', 'category'] for column in columns_to_convert_to_codes: df[column] = df[column].astype('category').cat.codes # Split the data into features and target X = df.drop(columns=['number of goals team2']) y = df['number of goals team2'] # Train the model X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=1) rfc_tuned = RandomForestClassifier(n_estimators=100, max_depth=10) rfc_tuned.fit(X_train, y_train) # Calculate feature importance importance_df = pd.DataFrame({"Feature_Name": X.columns, "Importance": rfc_tuned.feature_importances_}) sorted_importance_df = importance_df.sort_values(by="Importance", ascending=False) # Display feature importance st.subheader("Feature Importance") st.write("This chart shows the importance of each feature in predicting the number of goals scored by Team 2.") chart = st.bar_chart(sorted_importance_df.set_index('Feature_Name').head(15), use_container_width=True) # Explanation of feature importance st.subheader("Interpretation of Feature Importance") st.write("Feature importance indicates how much each feature influences the prediction.") st.write("Higher importance suggests stronger influence on predicting the number of goals.") # Create SHAP explainer explainer = shap.TreeExplainer(rfc_tuned) # Generate SHAP values shap_values = explainer.shap_values(X_test) # Display SHAP summary plot st.subheader("SHAP Values") st.write("SHAP values reveal the impact of each feature on individual predictions.") fig, ax = plt.subplots() shap.summary_plot(shap_values, X_test, plot_type='bar', max_display=10, show=False) st.pyplot(fig) # Toggle button to switch between feature importance and SHAP values show_feature_importance = st.checkbox("View Feature Importance Table") if show_feature_importance: st.write(sorted_importance_df.head(15)) # Slider to adjust number of features displayed in feature importance chart num_features = st.slider("Number of Features to Display", min_value=5, max_value=len(sorted_importance_df), value=10) st.bar_chart(sorted_importance_df.set_index('Feature_Name').head(num_features), use_container_width=True) # Explanation of SHAP values st.subheader("Interpretation of SHAP Values") st.write("Positive SHAP values indicate features that increase the predicted number of goals.") st.write("Negative SHAP values indicate features that decrease the predicted number of goals.") st.write("Higher magnitude suggests stronger impact on predictions.") elif st.session_state.app_mode == 'MLflow & Deployment': from sklearn.model_selection import train_test_split, GridSearchCV from sklearn.tree import DecisionTreeClassifier from sklearn import metrics from mlflow import log_metric import mlflow import os st.subheader("MLflow & Deployment") st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you! πΆ Feel free to adjust the volume or stop the song whenever you want. π§
", unsafe_allow_html=True) st.sidebar.audio(audio_fifa_4, format='audio/mp3', start_time=0) st.image(image_quatar2022_2, use_column_width=True) st.title("MLflow & Deployment") df = pd.read_csv('FIFAWorldCup2022.csv') X = df[["assists team2", "attempts inside the penalty area team2"]] # Features y = df['number of goals team2'] # Target X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42) dt = DecisionTreeClassifier(random_state=42) param_grid = {'max_depth': [3, 5, 10], 'min_samples_leaf': [1, 2, 4]} grid_search = GridSearchCV(estimator=dt, param_grid=param_grid, cv=5) grid_search.fit(X_train, y_train) best_params = grid_search.best_params_ mlflow.log_params(best_params) best_dt = grid_search.best_estimator_ y_pred = best_dt.predict(X_test) accuracy = metrics.accuracy_score(y_test, y_pred) precision = metrics.precision_score(y_test, y_pred, average='macro') recall = metrics.recall_score(y_test, y_pred, average='macro') f1 = metrics.f1_score(y_test, y_pred, average='macro') log_metric("accuracy", accuracy) log_metric("precision", precision) log_metric("recall", recall) log_metric("f1", f1) mlflow.sklearn.log_model(best_dt, "best") model_path = "best_model" if os.path.exists(model_path): try: import shutil shutil.rmtree(model_path) except OSError as e: st.error(f"An error occurred while deleting the previous model: {e}") mlflow.sklearn.save_model(best_dt, model_path) st.subheader("Performance Metrics:") fig, ax = plt.subplots(figsize=(10, 6)) metrics_names = ['Accuracy', 'Precision', 'Recall', 'F1 Score'] metrics_values = [accuracy, precision, recall, f1] bars = ax.bar(metrics_names, metrics_values, color=['blue', 'green', 'orange', 'red']) ax.set_ylabel('Score') ax.set_title('Performance Metrics') for bar in bars: height = bar.get_height() ax.annotate(f'{height:.3f}', xy=(bar.get_x() + bar.get_width() / 2, height), xytext=(0, 3), textcoords="offset points", ha='center', va='bottom', fontsize=12) st.pyplot(fig) # Assuming metrics_values and metrics_names are defined elsewhere metrics_values = [25, 35, 20, 20] metrics_names = ['Metric A', 'Metric B', 'Metric C', 'Metric D'] st.subheader("Additional Visualization (Pie Chart):") st.write("The pie chart illustrates the distribution of performance metrics.") # Create the pie chart with custom colors fig, ax = plt.subplots(figsize=(8, 6)) wedges, texts, autotexts = ax.pie(metrics_values, labels=metrics_names, autopct='%1.1f%%', startangle=140, colors=['blue', 'green', 'orange', 'red'], wedgeprops=dict(width=0.4)) # Set the color of autopct text to black for better visibility plt.setp(autotexts, size=12, weight="bold", color="black") # Adjust pie chart properties ax.axis('equal') ax.set_title('Performance Metrics Distribution') # Add values next to pie chart slices for i, text in enumerate(texts): text.set_text(f'{metrics_names[i]}: {metrics_values[i]:.3f}') # Display the pie chart st.pyplot(fig) st.info("Hover over the bars in the bar graph to view exact values. Click on the pie chart segments to see percentage breakdown.") st.subheader("Additional Insights:") st.write("Let's dive deeper into the performance metrics to understand their significance:") st.write("- **Accuracy**: Indicates the overall correctness of the model's predictions. A higher accuracy suggests better performance.") st.write("- **Precision**: Measures the correctness of positive predictions. It's the ratio of true positive predictions to all positive predictions made by the model.") st.write("- **Recall**: Reflects the model's ability to find all positive samples. It's the ratio of true positive predictions to all actual positive samples.") st.write("- **F1 Score**: Harmonic mean of precision and recall. It provides a balance between precision and recall, especially when dealing with imbalanced datasets.") import streamlit as st # Define questions and answers questions = [ { "question": "Which country won the first ever FIFA World Cup in 1930?", "options": ["", "Brazil", "Uruguay", "Argentina", "Italy"], "answer": "Uruguay" }, { "question": "Who is the all-time leading goal scorer in FIFA World Cup history?", "options": ["", "Pele", "Miroslav Klose", "Lionel Messi", "Cristiano Ronaldo"], "answer": "Miroslav Klose" } ] # Define congratulatory message congrats_message = "π Congratulations! You got it right! π" # Define function to display question and options def display_question(question_obj): st.subheader(question_obj["question"]) selected_option = st.radio("Select an option:", options=question_obj["options"]) if selected_option == question_obj["answer"] and selected_option != "": st.success(congrats_message) elif selected_option != "": st.warning("Oops! That's not correct. Keep trying!") # Prediction Page with Mindrefreshing Feature if st.session_state.app_mode == 'MLflow & Deployment': st.title("Mindrefreshing Feature: FIFA World Cup Trivia") st.markdown("Test your knowledge with these fun FIFA World Cup trivia questions!") # Display questions and options for i, question in enumerate(questions, 1): st.write(f"**Question {i}:**") display_question(question) # Option to play again for each question play_again = st.button("Play Again", key=f"play_again_{i}") if play_again: # Reset session state to reload questions st.session_state.app_mode = 'MLflow & Deployment' # Conclusion Page elif st.session_state.app_mode == 'Conclusion': st.subheader("Conclusion") # Play the FIFA song st.sidebar.subheader("Play FIFA World Cup Song") st.sidebar.markdown("π΅ Enjoy the below chosen FIFA World Cup song for you! πΆ Feel free to adjust the volume or stop the song whenever you want. π§
", unsafe_allow_html=True) st.sidebar.audio(audio_fifa, format='audio/mp3') st.title("FIFA World Cup 2022 Data Analysis - Conclusion π") st.video(video_concu, format='video/mp4') # Set page background color and font st.markdown( """ """, unsafe_allow_html=True ) # Insights about team performance st.markdown("## Team Performance Insights") st.markdown("1. **Accuracy Matters:** Teams with precise shots tend to score more goals.") st.markdown("2. **Seize the Opportunities:** More shots on target often translate to more scoring chances.") st.markdown("3. **Balancing Act:** Teams that excel in attack also need to maintain a solid defense.") st.markdown("4. **Team Play:** Assists play a crucial role in achieving higher goal counts.") st.markdown("5. **Defensive Tactics:** Aggressive defensive strategies can lead to fewer goals conceded.") # Limitations st.markdown("## Limitations") st.markdown("1. **Correlation, Not Causation:** While our models show strong correlations, causation cannot be definitively claimed.") st.markdown("2. **Room for Improvement:** Our prediction models require refinement for greater accuracy.") st.markdown("3. **Work in Progress:** Currently, our analysis does not predict game winners, as it wasn't our primary focus.") # Future directions st.markdown("## Future Directions") st.markdown("1. **Time Is Key:** Investigate the impact of specific game minutes on goal likelihood in real-time.") st.markdown("2. **Beyond the Numbers:** Explore sentiment analysis to understand player and fan dynamics and their influence on goals.") st.markdown("3. **Stay Updated:** Implement real-time data analysis for timely insights during tournaments.") st.markdown("4. **Enhanced Predictions:** Develop robust models based on historical data to predict match outcomes and winners.") import streamlit as st import pandas as pd import numpy as np from sklearn.model_selection import train_test_split from sklearn.linear_model import LinearRegression from sklearn.metrics import mean_squared_error from codecarbon import EmissionsTracker import tensorflow as tf # Load FIFA World Cup dataset from CSV data = pd.read_csv('FIFAWorldCup2022.csv') # Select independent variables (features) and target variable selected_features = ["assists team2", "attempted defensive line breaks team2", "on target attempts team2", "inbehind offers to receive team2", "possession team2", "passes completed team2"] selected_target = 'number of goals team2' # Remove percentage signs and convert to float for column in selected_features: data[column] = data[column].astype(str).str.rstrip('%').astype(float) # Extract selected features and target variable from the dataset X = data[selected_features] y = data[selected_target] # 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 the emissions tracker for linear regression tracker_linear = EmissionsTracker() tracker_linear.start() # Train the linear regression model model_linear = LinearRegression() model_linear.fit(X_train, y_train) # Predict the house prices predictions_linear = model_linear.predict(X_test) # Stop the emissions tracker for linear regression emissions_linear = tracker_linear.stop() print(f"Estimated emissions for training the linear regression model: {emissions_linear:.4f} kg of CO2") # Evaluate the linear regression model mse_linear = mean_squared_error(y_test, predictions_linear) rmse_linear = np.sqrt(mse_linear) print("Root Mean Squared Error (Linear Regression):", rmse_linear) # Define a function to load MNIST dataset def load_mnist(): mnist = tf.keras.datasets.mnist (x_train, y_train), (x_test, y_test) = mnist.load_data() x_train, x_test = x_train / 255.0, x_test / 255.0 return (x_train, y_train), (x_test, y_test) # Load MNIST dataset (x_train, y_train), (x_test, y_test) = load_mnist() # Initialize the emissions tracker for neural network tracker_nn = EmissionsTracker() tracker_nn.start() # Define and train the neural network model model_nn = tf.keras.models.Sequential([ tf.keras.layers.Flatten(input_shape=(28, 28)), tf.keras.layers.Dense(128, activation="relu"), tf.keras.layers.Dropout(0.2), tf.keras.layers.Dense(10) ]) loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True) model_nn.compile(optimizer="adam", loss=loss_fn, metrics=["accuracy"]) model_nn.fit(x_train, y_train, epochs=3) # Stop the emissions tracker for neural network emissions_nn = tracker_nn.stop() print(f"Estimated emissions for training the neural network model: {emissions_nn:.4f} kg of CO2") # Combine emissions from both models total_emissions = emissions_linear + emissions_nn # Calculate accuracy of the neural network model test_loss, test_accuracy = model_nn.evaluate(x_test, y_test, verbose=2) print("Test Accuracy (Neural Network):", test_accuracy) # Button to toggle the visibility of the output if st.button("Show Emissions and Model Evaluation"): # Estimated emissions and model evaluation st.markdown("## Model Evaluation and Environmental Impact") st.markdown("A. Estimated emissions for training the linear regression model:") st.write(f"{emissions_linear:.4f} kg of CO2") st.markdown("B. Root Mean Squared Error (Linear Regression):") st.write(rmse_linear) st.markdown("C. Estimated emissions for training the neural network model:") st.write(f"{emissions_nn:.4f} kg of CO2") st.markdown("D. Total emissions:") st.write(f"{total_emissions:.4f} kg of CO2") st.markdown("E. Test Accuracy (Neural Network):") st.write(test_accuracy) # Display questions below emissions button st.markdown("## Kahoot Quiz") questions = [ { "question": "Which of the following is a key component that increases the likelihood of a team scoring goals?", "options": ["", "On Target Attempts", "Number of Fans in the Stadium", "Weather Conditions", "Team's Mascot"], "answer": "On Target Attempts", "selected_option": None }, { "question": "Which factor is most crucial for a team to create scoring opportunities?", "options": ["", "Number of Goals Conceded", "Successful Passes Completed by the Team", "Team's Jersey Color", "Length of the Grass on the Field"], "answer": "Successful Passes Completed by the Team", "selected_option": None } ] congrats_message = "π Congratulations! You got it right! π" # Define function to display question and options def display_question(question_obj): st.markdown(f"### {question_obj['question']}") selected_option = st.radio("Select your answer:", options=question_obj["options"], key=question_obj["question"]) if selected_option == question_obj["answer"]: st.success(congrats_message) elif selected_option and selected_option != "": st.warning("Oops! That's not correct. Better luck next time!") # Quiz for i, question in enumerate(questions, 1): st.write(f"**Question {i}:**") display_question(question) # Conclusion and Surprise Element st.markdown("## π That's a Wrap! π
", unsafe_allow_html=True) st.markdown("π Thanks for exploring our FIFA World Cup 2022 Data Analysis app! π
", unsafe_allow_html=True) st.markdown("Hope you enjoyed discovering insights and trends in the data.
", unsafe_allow_html=True) st.markdown("Congratulations on your journey through football analytics!
", unsafe_allow_html=True) st.markdown("Here's a special surprise just for you!
", unsafe_allow_html=True) # Additional Shocking Feature st.subheader("Reveal Secret") if st.button("Reveal Secret"): st.balloons() st.success("π You found the hidden treasure! Enjoy your victory! π") # Advanced Feature: Continuously moving and shining balloons st.write('', unsafe_allow_html=True) st.write('', unsafe_allow_html=True)