sqlshell 0.4.4__py3-none-any.whl

sqlshell/utils/profile_distributions.py

@@ -0,0 +1,616 @@
+import pandas as pd
+import numpy as np
+import scipy.stats as stats
+import matplotlib
+try:
+    matplotlib.use('qtagg')  # Set the backend before importing pyplot
+except ImportError:
+    matplotlib.use('Agg')  # Fall back to headless backend for CI/testing
+import matplotlib.pyplot as plt
+from matplotlib.figure import Figure
+from matplotlib.backends.backend_qtagg import FigureCanvasQTAgg
+from PyQt6.QtCore import QObject, pyqtSignal, Qt
+from PyQt6.QtWidgets import (
+    QApplication, QMainWindow, QVBoxLayout, QHBoxLayout, QWidget, 
+    QTableView, QHeaderView, QLabel, QFrame, QScrollArea, QTabWidget,
+    QComboBox, QPushButton, QSplitter
+)
+from PyQt6.QtGui import QStandardItemModel, QStandardItem, QColor, QBrush
+
+
+class DistributionProfiler(QObject):
+    """Class to analyze distributions of columns in a dataframe"""
+    progress_updated = pyqtSignal(int, str)  # Signal for progress reporting
+    
+    def __init__(self):
+        super().__init__()
+        
+        # Define common distributions to test
+        self.distributions = [
+            {'name': 'normal', 'distribution': stats.norm, 'color': 'blue'},
+            {'name': 'uniform', 'distribution': stats.uniform, 'color': 'green'},
+            {'name': 'exponential', 'distribution': stats.expon, 'color': 'red'},
+            {'name': 'lognormal', 'distribution': stats.lognorm, 'color': 'purple'},
+            {'name': 'gamma', 'distribution': stats.gamma, 'color': 'orange'},
+            {'name': 'beta', 'distribution': stats.beta, 'color': 'brown'},
+        ]
+    
+    def get_best_distribution(self, data):
+        """Find the best distribution that fits the data"""
+        # Remove NaNs
+        data = data.dropna()
+        
+        if len(data) == 0:
+            return None, None, None
+            
+        # For categorical or non-numeric data, return None
+        if not pd.api.types.is_numeric_dtype(data):
+            return None, None, None
+            
+        # For constant data, return a simple result
+        if data.nunique() == 1:
+            return 'constant', None, 1.0
+        
+        # If too few unique values, may not be appropriate for distribution fitting
+        if data.nunique() < 5:
+            return 'discrete', None, None
+        
+        best_distribution = None
+        best_params = None
+        best_sse = np.inf
+        
+        # Get histogram data
+        hist, bin_edges = np.histogram(data, bins='auto', density=True)
+        bin_centers = (bin_edges[:-1] + bin_edges[1:]) / 2
+        
+        # Try each distribution
+        for dist_info in self.distributions:
+            distribution = dist_info['distribution']
+            
+            try:
+                # Fit distribution to data
+                params = distribution.fit(data)
+                
+                # Get PDF values
+                arg_params = params[:-2]
+                loc = params[-2]
+                scale = params[-1]
+                
+                pdf = distribution.pdf(bin_centers, loc=loc, scale=scale, *arg_params)
+                
+                # Calculate sum of squared errors
+                sse = np.sum((pdf - hist) ** 2)
+                
+                # Find best fit distribution
+                if sse < best_sse:
+                    best_distribution = dist_info['name']
+                    best_params = params
+                    best_sse = sse
+                    
+            except Exception:
+                continue
+        
+        # Calculate Kolmogorov-Smirnov test for goodness of fit
+        if best_distribution and best_params:
+            dist = getattr(stats, best_distribution)
+            
+            # Try to compute K-S test
+            try:
+                arg_params = best_params[:-2]
+                loc = best_params[-2]
+                scale = best_params[-1]
+                ks_stat, p_value = stats.kstest(data, dist.cdf, args=arg_params, loc=loc, scale=scale)
+                return best_distribution, best_params, p_value
+            except:
+                return best_distribution, best_params, None
+        
+        return None, None, None
+    
+    def describe_distribution(self, series):
+        """Provide distribution statistics for a series"""
+        stats_dict = {}
+        
+        # Remove NaNs
+        series = series.dropna()
+        
+        if len(series) == 0:
+            return {
+                'count': 0,
+                'distribution': 'empty',
+                'goodness_of_fit': None
+            }
+        
+        # Basic statistics
+        stats_dict['count'] = len(series)
+        stats_dict['unique_count'] = series.nunique()
+        stats_dict['missing_count'] = series.isna().sum()
+        stats_dict['missing_percentage'] = (series.isna().sum() / len(series)) * 100
+        
+        # For categorical data
+        if not pd.api.types.is_numeric_dtype(series):
+            stats_dict['type'] = 'categorical'
+            top_values = series.value_counts().head(5).to_dict()
+            stats_dict['top_values'] = {str(k): v for k, v in top_values.items()}
+            stats_dict['distribution'] = 'categorical'
+            return stats_dict
+        
+        # For numerical data
+        stats_dict['type'] = 'numerical'
+        stats_dict['min'] = float(series.min())
+        stats_dict['max'] = float(series.max())
+        stats_dict['mean'] = float(series.mean())
+        stats_dict['median'] = float(series.median())
+        stats_dict['std'] = float(series.std())
+        stats_dict['skewness'] = float(stats.skew(series))
+        stats_dict['kurtosis'] = float(stats.kurtosis(series))
+        
+        # Find best distribution
+        best_dist, params, p_value = self.get_best_distribution(series)
+        stats_dict['distribution'] = best_dist
+        stats_dict['goodness_of_fit'] = p_value
+        
+        return stats_dict
+    
+    def profile(self, df):
+        """
+        Profile a dataframe to identify the distribution characteristics of each column.
+        
+        Args:
+            df: pandas DataFrame to analyze
+            
+        Returns:
+            DataFrame with columns and their distribution profiles
+        """
+        if not isinstance(df, pd.DataFrame):
+            raise TypeError("Input must be a pandas DataFrame")
+        
+        if df.empty:
+            return pd.DataFrame(columns=['column', 'type', 'distribution', 'goodness_of_fit'])
+        
+        results = []
+        total_columns = len(df.columns)
+        
+        # Analyze each column
+        for i, column in enumerate(df.columns):
+            # Emit progress signal (if connected)
+            self.progress_updated.emit(int((i / total_columns) * 100), f"Analyzing column: {column}")
+            
+            try:
+                stats_dict = self.describe_distribution(df[column])
+                stats_dict['column'] = column
+                results.append(stats_dict)
+            except Exception as e:
+                # Skip columns that can't be analyzed
+                continue
+        
+        # Create results dataframe
+        result_df = pd.DataFrame(results)
+        
+        if result_df.empty:
+            return pd.DataFrame(columns=['column', 'type', 'distribution', 'goodness_of_fit'])
+        
+        # Sort by distribution type and column name
+        result_df = result_df.sort_values(by=['type', 'column'])
+        
+        self.progress_updated.emit(100, "Analysis complete")
+        return result_df
+
+
+class MatplotlibCanvas(FigureCanvasQTAgg):
+    """Matplotlib canvas for embedding plots in PyQt"""
+    def __init__(self, width=5, height=4, dpi=100):
+        self.fig = Figure(figsize=(width, height), dpi=dpi)
+        self.axes = self.fig.add_subplot(111)
+        super().__init__(self.fig)
+
+
+class DistributionVisualization(QMainWindow):
+    """Window to visualize distribution results"""
+    
+    def __init__(self, df, results_df, parent=None):
+        super().__init__(parent)
+        self.df = df
+        self.results_df = results_df
+        self.current_column = None
+        
+        self.setWindowTitle("Column Distribution Profiles")
+        self.resize(1000, 800)
+        
+        # Create central widget and layout
+        central_widget = QWidget()
+        self.setCentralWidget(central_widget)
+        main_layout = QVBoxLayout(central_widget)
+        
+        # Add a title
+        title = QLabel("Statistical Distribution Analysis")
+        title.setAlignment(Qt.AlignmentFlag.AlignCenter)
+        title.setStyleSheet("font-size: 16pt; font-weight: bold; margin: 10px;")
+        main_layout.addWidget(title)
+        
+        # Add a description
+        description = QLabel(
+            "Analyzing column distributions helps identify data patterns and select appropriate statistical methods."
+        )
+        description.setAlignment(Qt.AlignmentFlag.AlignCenter)
+        description.setWordWrap(True)
+        main_layout.addWidget(description)
+        
+        # Create a splitter for table and visualization
+        splitter = QSplitter(Qt.Orientation.Vertical)
+        main_layout.addWidget(splitter, 1)
+        
+        # Create table view
+        table_widget = QWidget()
+        table_layout = QVBoxLayout(table_widget)
+        self.create_table_view(table_layout)
+        splitter.addWidget(table_widget)
+        
+        # Create visualization section
+        vis_widget = QWidget()
+        vis_layout = QVBoxLayout(vis_widget)
+        self.create_visualization_section(vis_layout)
+        splitter.addWidget(vis_widget)
+        
+        # Set initial splitter sizes
+        splitter.setSizes([300, 500])
+    
+    def create_table_view(self, layout):
+        """Create a table view showing the distribution results"""
+        # Create the model
+        model = QStandardItemModel()
+        headers = ['Column', 'Type', 'Distribution', 'Count', 'Unique', 'Missing %']
+        if 'skewness' in self.results_df.columns:
+            headers.extend(['Mean', 'Median', 'Std', 'Skewness', 'Kurtosis'])
+        model.setHorizontalHeaderLabels(headers)
+        
+        # Set table data
+        for _, row in self.results_df.iterrows():
+            items = []
+            
+            # Basic columns present in all rows
+            column_item = QStandardItem(str(row['column']))
+            type_item = QStandardItem(str(row['type']))
+            dist_item = QStandardItem(str(row['distribution']))
+            count_item = QStandardItem(str(row['count']))
+            unique_item = QStandardItem(str(row.get('unique_count', 'N/A')))
+            missing_item = QStandardItem(f"{row.get('missing_percentage', 0):.1f}%")
+            
+            items.extend([column_item, type_item, dist_item, count_item, unique_item, missing_item])
+            
+            # Add numerical statistics if available
+            if row['type'] == 'numerical':
+                mean_item = QStandardItem(f"{row.get('mean', 'N/A'):.2f}")
+                median_item = QStandardItem(f"{row.get('median', 'N/A'):.2f}")
+                std_item = QStandardItem(f"{row.get('std', 'N/A'):.2f}")
+                skew_item = QStandardItem(f"{row.get('skewness', 'N/A'):.2f}")
+                kurt_item = QStandardItem(f"{row.get('kurtosis', 'N/A'):.2f}")
+                
+                items.extend([mean_item, median_item, std_item, skew_item, kurt_item])
+            else:
+                # Add empty items for categorical data
+                for _ in range(5):
+                    items.append(QStandardItem(""))
+            
+            model.appendRow(items)
+        
+        # Create and configure the table view
+        self.table_view = QTableView()
+        self.table_view.setModel(model)
+        self.table_view.horizontalHeader().setSectionResizeMode(QHeaderView.ResizeMode.ResizeToContents)
+        self.table_view.setAlternatingRowColors(True)
+        self.table_view.setSelectionBehavior(QTableView.SelectionBehavior.SelectRows)
+        self.table_view.setMinimumHeight(200)
+        
+        # Connect selection signal
+        self.table_view.selectionModel().selectionChanged.connect(self.on_column_selected)
+        
+        layout.addWidget(self.table_view)
+        
+        # Add column selector
+        selector_layout = QHBoxLayout()
+        selector_layout.addWidget(QLabel("Select Column:"))
+        
+        self.column_selector = QComboBox()
+        self.column_selector.addItems(self.results_df['column'].tolist())
+        self.column_selector.currentTextChanged.connect(self.on_combobox_changed)
+        selector_layout.addWidget(self.column_selector)
+        
+        layout.addLayout(selector_layout)
+    
+    def create_visualization_section(self, layout):
+        """Create the visualization section with tabs for different plots"""
+        self.tab_widget = QTabWidget()
+        
+        # Create tabs for different visualizations
+        self.histogram_tab = QWidget()
+        self.histogram_layout = QVBoxLayout(self.histogram_tab)
+        self.histogram_canvas = MatplotlibCanvas(width=8, height=4)
+        self.histogram_layout.addWidget(self.histogram_canvas)
+        self.tab_widget.addTab(self.histogram_tab, "Histogram & Density")
+        
+        self.boxplot_tab = QWidget()
+        self.boxplot_layout = QVBoxLayout(self.boxplot_tab)
+        self.boxplot_canvas = MatplotlibCanvas(width=8, height=4)
+        self.boxplot_layout.addWidget(self.boxplot_canvas)
+        self.tab_widget.addTab(self.boxplot_tab, "Box Plot")
+        
+        self.qq_tab = QWidget()
+        self.qq_layout = QVBoxLayout(self.qq_tab)
+        self.qq_canvas = MatplotlibCanvas(width=8, height=4)
+        self.qq_layout.addWidget(self.qq_canvas)
+        self.tab_widget.addTab(self.qq_tab, "Q-Q Plot")
+        
+        self.ecdf_tab = QWidget()
+        self.ecdf_layout = QVBoxLayout(self.ecdf_tab)
+        self.ecdf_canvas = MatplotlibCanvas(width=8, height=4)
+        self.ecdf_layout.addWidget(self.ecdf_canvas)
+        self.tab_widget.addTab(self.ecdf_tab, "Empirical CDF")
+        
+        # For categorical data
+        self.categorical_tab = QWidget()
+        self.categorical_layout = QVBoxLayout(self.categorical_tab)
+        self.categorical_canvas = MatplotlibCanvas(width=8, height=4)
+        self.categorical_layout.addWidget(self.categorical_canvas)
+        self.tab_widget.addTab(self.categorical_tab, "Bar Chart")
+        
+        layout.addWidget(self.tab_widget)
+        
+        # Stats panel
+        self.stats_label = QLabel("Select a column to view distribution statistics")
+        self.stats_label.setAlignment(Qt.AlignmentFlag.AlignCenter)
+        self.stats_label.setStyleSheet("font-family: monospace; background-color: #f0f0f0; padding: 10px; border-radius: 5px;")
+        self.stats_label.setWordWrap(True)
+        layout.addWidget(self.stats_label)
+    
+    def on_combobox_changed(self, column_name):
+        """Handle column selection from combobox"""
+        self.visualize_column(column_name)
+    
+    def on_column_selected(self, selected, deselected):
+        """Handle column selection from table"""
+        indexes = selected.indexes()
+        if indexes:
+            # Get the column name from the first column
+            row_idx = indexes[0].row()
+            column_name = self.table_view.model().item(row_idx, 0).text()
+            
+            # Update combobox to match
+            index = self.column_selector.findText(column_name)
+            if index >= 0:
+                self.column_selector.setCurrentIndex(index)
+            
+            self.visualize_column(column_name)
+    
+    def visualize_column(self, column_name):
+        """Visualize the selected column with various plots"""
+        if column_name not in self.df.columns:
+            return
+            
+        self.current_column = column_name
+        
+        # Get column data and stats
+        series = self.df[column_name]
+        column_stats = self.results_df[self.results_df['column'] == column_name].iloc[0].to_dict()
+        
+        # Update stats label
+        self.update_stats_display(column_stats)
+        
+        # Check if categorical or numerical
+        if column_stats['type'] == 'categorical':
+            self.create_categorical_plots(series, column_stats)
+            self.tab_widget.setCurrentWidget(self.categorical_tab)
+        else:
+            self.create_numerical_plots(series, column_stats)
+            self.tab_widget.setCurrentWidget(self.histogram_tab)
+    
+    def update_stats_display(self, stats):
+        """Update the statistics display panel"""
+        if stats['type'] == 'numerical':
+            # Format numerical stats
+            stats_text = (
+                f"<b>Column:</b> {stats['column']} | <b>Type:</b> {stats['type']} | "
+                f"<b>Distribution:</b> {stats['distribution']}\n"
+                f"<b>Count:</b> {stats['count']} | <b>Unique:</b> {stats['unique_count']} | "
+                f"<b>Missing:</b> {stats['missing_count']} ({stats['missing_percentage']:.1f}%)\n"
+                f"<b>Min:</b> {stats['min']:.4g} | <b>Max:</b> {stats['max']:.4g} | "
+                f"<b>Mean:</b> {stats['mean']:.4g} | <b>Median:</b> {stats['median']:.4g} | <b>Std:</b> {stats['std']:.4g}\n"
+                f"<b>Skewness:</b> {stats['skewness']:.4g} | <b>Kurtosis:</b> {stats['kurtosis']:.4g}"
+            )
+            
+            if stats['goodness_of_fit'] is not None:
+                stats_text += f" | <b>Goodness of fit (p-value):</b> {stats['goodness_of_fit']:.4g}"
+        else:
+            # Format categorical stats
+            stats_text = (
+                f"<b>Column:</b> {stats['column']} | <b>Type:</b> {stats['type']}\n"
+                f"<b>Count:</b> {stats['count']} | <b>Unique:</b> {stats['unique_count']} | "
+                f"<b>Missing:</b> {stats['missing_count']} ({stats['missing_percentage']:.1f}%)"
+            )
+            
+            if 'top_values' in stats:
+                top_values = stats['top_values']
+                stats_text += "\n<b>Top values:</b> "
+                stats_text += ", ".join([f"{k} ({v})" for k, v in top_values.items()])
+        
+        self.stats_label.setText(stats_text)
+    
+    def create_numerical_plots(self, series, column_stats):
+        """Create plots for numerical data"""
+        # Clean data
+        data = series.dropna()
+        
+        # Histogram with fitted distribution
+        self.histogram_canvas.axes.clear()
+        self.histogram_canvas.axes.hist(data, bins='auto', density=True, alpha=0.6, label="Data")
+        
+        # If we have a fitted distribution, plot it
+        if column_stats['distribution'] not in [None, 'discrete', 'constant', 'categorical']:
+            # Get the distribution and params
+            dist_name = column_stats['distribution']
+            
+            # Simple estimation for distribution parameters if we don't have them
+            # In a real implementation, you would save the parameters from the profiler
+            if dist_name == 'normal':
+                x = np.linspace(data.min(), data.max(), 1000)
+                y = stats.norm.pdf(x, data.mean(), data.std())
+                self.histogram_canvas.axes.plot(x, y, 'r-', lw=2, label=f"Fitted {dist_name}")
+            elif dist_name == 'uniform':
+                x = np.linspace(data.min(), data.max(), 1000)
+                y = stats.uniform.pdf(x, data.min(), data.max() - data.min())
+                self.histogram_canvas.axes.plot(x, y, 'r-', lw=2, label=f"Fitted {dist_name}")
+            elif dist_name == 'exponential':
+                x = np.linspace(data.min(), data.max(), 1000)
+                y = stats.expon.pdf(x, scale=1/data.mean())
+                self.histogram_canvas.axes.plot(x, y, 'r-', lw=2, label=f"Fitted {dist_name}")
+            
+        self.histogram_canvas.axes.set_title(f"Histogram of {series.name}")
+        self.histogram_canvas.axes.set_xlabel("Value")
+        self.histogram_canvas.axes.set_ylabel("Density")
+        self.histogram_canvas.axes.legend()
+        self.histogram_canvas.fig.tight_layout()
+        self.histogram_canvas.draw()
+        
+        # Box plot
+        self.boxplot_canvas.axes.clear()
+        self.boxplot_canvas.axes.boxplot(data, vert=False)
+        self.boxplot_canvas.axes.set_title(f"Box Plot of {series.name}")
+        self.boxplot_canvas.axes.set_xlabel("Value")
+        self.boxplot_canvas.axes.set_yticks([])
+        self.boxplot_canvas.fig.tight_layout()
+        self.boxplot_canvas.draw()
+        
+        # Q-Q plot
+        self.qq_canvas.axes.clear()
+        stats.probplot(data, dist="norm", plot=self.qq_canvas.axes)
+        self.qq_canvas.axes.set_title(f"Q-Q Plot of {series.name} (vs Normal)")
+        self.qq_canvas.fig.tight_layout()
+        self.qq_canvas.draw()
+        
+        # Empirical CDF
+        self.ecdf_canvas.axes.clear()
+        x = np.sort(data)
+        y = np.arange(1, len(x) + 1) / len(x)
+        self.ecdf_canvas.axes.step(x, y, where='post', label="Empirical CDF")
+        self.ecdf_canvas.axes.set_title(f"Empirical CDF of {series.name}")
+        self.ecdf_canvas.axes.set_xlabel("Value")
+        self.ecdf_canvas.axes.set_ylabel("Cumulative Probability")
+        self.ecdf_canvas.fig.tight_layout()
+        self.ecdf_canvas.draw()
+    
+    def create_categorical_plots(self, series, stats):
+        """Create plots for categorical data"""
+        # Clean data
+        data = series.dropna()
+        
+        # Bar chart for categorical data
+        self.categorical_canvas.axes.clear()
+        value_counts = data.value_counts().sort_values(ascending=False)
+        
+        # Limit to top 15 categories if there are too many
+        if len(value_counts) > 15:
+            value_counts = value_counts.head(15)
+            title = f"Top 15 Categories in {series.name}"
+        else:
+            title = f"Categories in {series.name}"
+            
+        value_counts.plot(kind='bar', ax=self.categorical_canvas.axes)
+        self.categorical_canvas.axes.set_title(title)
+        self.categorical_canvas.axes.set_xlabel("Category")
+        self.categorical_canvas.axes.set_ylabel("Count")
+        
+        # Rotate x-axis labels if needed
+        if len(value_counts) > 5:
+            plt.setp(self.categorical_canvas.axes.get_xticklabels(), rotation=45, ha='right')
+            
+        self.categorical_canvas.fig.tight_layout()
+        self.categorical_canvas.draw()
+
+
+# Function interface for simpler usage
+def profile(df):
+    """
+    Profile a dataframe to identify the distribution characteristics of each column.
+    
+    Args:
+        df: pandas DataFrame to analyze
+        
+    Returns:
+        DataFrame with columns and their distribution profiles
+    """
+    profiler = DistributionProfiler()
+    return profiler.profile(df)
+
+
+def visualize_profile(df):
+    """
+    Create a visual representation of the distribution profiles for a dataframe.
+    
+    Args:
+        df: pandas DataFrame to analyze
+        
+    Returns:
+        A PyQt6 window showing the visualization
+    """
+    profiler = DistributionProfiler()
+    results = profiler.profile(df)
+    vis = DistributionVisualization(df, results)
+    vis.show()
+    return vis
+
+
+def test_profile_distributions():
+    """Test the distribution profiler with a sample dataframe"""
+    import sys
+    
+    # Create a QApplication instance if one doesn't exist
+    app = QApplication.instance()
+    if app is None:
+        app = QApplication(sys.argv)
+    
+    # Generate a random dataframe with some columns with different distributions
+    np.random.seed(42)  # For reproducibility
+    
+    # Create a dataframe with columns of varying distributions
+    df = pd.DataFrame({
+        'uniform': np.random.uniform(0, 100, size=1000),  # Uniform distribution
+        'normal': np.random.normal(50, 10, size=1000),    # Normal distribution
+        'exponential': np.random.exponential(5, size=1000),  # Exponential distribution
+        'lognormal': np.random.lognormal(0, 1, size=1000),  # Log-normal distribution
+        'bimodal': np.concatenate([np.random.normal(20, 5, 500), np.random.normal(60, 5, 500)]),  # Bimodal
+        'constant': np.ones(1000),  # Constant value
+        'binary': np.random.choice([0, 1], size=1000),  # Binary
+        'categorical': np.random.choice(['A', 'B', 'C', 'D'], size=1000),  # Categorical data
+        'skewed': 100 - np.random.power(5, size=1000) * 100,  # Right-skewed
+        'multimodal': np.concatenate([
+            np.random.normal(10, 2, 300),
+            np.random.normal(30, 2, 300),
+            np.random.normal(50, 2, 400)
+        ]),  # Multimodal
+        'boolean': np.random.choice([True, False], size=1000),  # Boolean
+        'integer': np.random.randint(1, 10, size=1000),  # Small integers
+        'text': pd.Series(['short', 'medium length', 'very long text entry', 'another value'] * 250)  # Text data
+    })
+    
+    # Add datetime and timedelta columns
+    df['datetime'] = pd.date_range('2020-01-01', periods=1000, freq='h')
+    df['timedelta'] = pd.Series([pd.Timedelta(days=i) for i in range(1000)])
+    
+    # Add a column with missing values
+    df['with_missing'] = df['normal'].copy()
+    df.loc[np.random.choice(df.index, size=200), 'with_missing'] = np.nan
+    
+    # Calculate and display profile information
+    print("Distribution Profile Results:")
+    profiler = DistributionProfiler()
+    result = profiler.profile(df)
+    print(result[['column', 'type', 'distribution', 'unique_count']])
+    
+    # Visualize the results
+    vis = visualize_profile(df)
+    
+    # Start the application event loop
+    app.exec()
+
+
+if __name__ == "__main__":
+    test_profile_distributions()