Update app.py

Adding skew ratio & momentum, cleaning up the visual.

app.py

@@ -90,7 +90,7 @@ def load_columns(file):
 def update_ma_visibility(add_ma):
     return gr.Slider(visible=add_ma)
 
-def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_length, confidence, add_trendline, add_moving_average, ma_window):
+def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_length, confidence, add_trendline, add_moving_average, ma_window, add_skew_viz):
     if file is None or time_col is None or selected_col is None:
         return None, "### Error\nPlease upload a CSV and select time and value columns!"
 
@@ -162,8 +162,11 @@ def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_le
         lower_slider = np.zeros(pred_len)
         upper_slider = np.zeros(pred_len)
 
+        skew_ratios = np.zeros(pred_len)
+        delta_skews = np.zeros(pred_len)
         skew_directions = []
 
+        epsilon = 1e-8
         for t in range(pred_len):
             q_t = q[t]
             lower50[t] = np.interp(lower_alpha_50, alphas, q_t)
@@ -175,9 +178,19 @@ def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_le
             med = median[t]
             upside_dist = upper_slider[t] - med
             downside_dist = med - lower_slider[t]
-            if upside_dist > downside_dist:
+            total_dist = upside_dist + downside_dist + epsilon
+            skew_ratios[t] = (upside_dist - downside_dist) / total_dist
+            
+            # Delta for momentum (shift from previous step)
+            if t == 0:
+                delta_skews[t] = 0.0
+            else:
+                delta_skews[t] = skew_ratios[t] - skew_ratios[t-1]
+            
+            # Existing categorical (optional: derive from skew_ratio for compat)
+            if skew_ratios[t] > 0.1:
                 skew_directions.append("Upside")
-            elif downside_dist > upside_dist:
+            elif skew_ratios[t] < -0.1:
                 skew_directions.append("Downside")
             else:
                 skew_directions.append("Neutral")
@@ -198,35 +211,44 @@ def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_le
             'predicted_sales_lower': lower_slider,
             'predicted_sales_upper': upper_slider,
             'predicted_sales_mean': mean_forecast,
-            'skew_direction': skew_directions
+            'skew_direction': skew_directions,
+            'skew_ratio': skew_ratios,
+            'delta_skew': delta_skews
         }).set_index('date')
 
         # Count skews for summary
-        upside_count = skew_directions.count("Upside")
-        downside_count = skew_directions.count("Downside")
-        neutral_count = skew_directions.count("Neutral")
+        upside_count = sum(1 for r in skew_ratios if r > 0.1)
+        downside_count = sum(1 for r in skew_ratios if r < -0.1)
+        neutral_count = pred_len - upside_count - downside_count
+
+        # NEW: Summary stats for skew momentum
+        avg_skew = skew_ratios.mean()
+        max_momentum_shift = abs(delta_skews).max()
 
         # Prepare markdown output (broken into smaller strings to avoid multiline f-string parsing issues)
         markdown_text = "### Summary\n"
         markdown_text += "- **Number of Historical Periods Used:** {} points\n".format(len(context_series))
         markdown_text += "- **Held Out Periods:** {} points {}\n".format(len(held_out_df), "(Full Context Used)" if len(held_out_df) == 0 else "(For Validation)")
         markdown_text += "- **Prediction Length:** {} periods\n".format(pred_len)
-        markdown_text += "- **Confidence Level:** {}% (alphas: {:.3f} - {:.3f})\n".format(confidence, lower_alpha_slider, upper_alpha_slider)
+        markdown_text += "- **Prediction Interval:** {}% (alphas: {:.3f} - {:.3f})\n".format(confidence, lower_alpha_slider, upper_alpha_slider)
         markdown_text += "- **Sum of Median Predicted Values:** {:.2f}\n".format(pred_df['predicted_sales_median'].sum())
         markdown_text += "- **Sum of Mean Predicted Values:** {:.2f}\n".format(pred_df['predicted_sales_mean'].sum())
-        markdown_text += "- **Skew Distribution:** {} Upside, {} Downside, {} Neutral\n\n".format(upside_count, downside_count, neutral_count)
+        markdown_text += "- **Skew Distribution:** {} Upside, {} Downside, {} Neutral\n".format(upside_count, downside_count, neutral_count)
+        markdown_text += "- **Average Skew Ratio:** {:.3f} (momentum: max |Δ| = {:.3f})\n\n".format(avg_skew, max_momentum_shift)
 
-        forecast_table = "### TiRex Forecast Results (Median + {}% Interval)\n\n".format(confidence)
-        forecast_table += "| Date | Median | Lower Bound | Upper Bound | Mean | Skew |\n"
-        forecast_table += "|------|--------|-------------|-------------|------|------|\n"
+        forecast_table = "### TiRex Forecast Results (Median + {}% Prediction Interval)\n\n".format(confidence)
+        forecast_table += "| Date | Median | Lower Bound | Upper Bound | Mean | Skew Direction | Skew Ratio | Δ Skew |\n"
+        forecast_table += "|------|--------|-------------|-------------|------|----------------|------------|--------|\n"
         for idx, row in pred_df.iterrows():
-            forecast_table += "| {} | {:.2f} | {:.2f} | {:.2f} | {:.2f} | {} |\n".format(
+            forecast_table += "| {} | {:.2f} | {:.2f} | {:.2f} | {:.2f} | {} | {:.3f} | {:.3f} |\n".format(
                 idx.strftime('%Y-%m-%d'),
                 row['predicted_sales_median'],
                 row['predicted_sales_lower'],
                 row['predicted_sales_upper'],
                 row['predicted_sales_mean'],
-                row['skew_direction']
+                row['skew_direction'],
+                row['skew_ratio'],
+                row['delta_skew']
             )
 
         sample_data = "### Sample Historical Data (Context)\n"
@@ -235,9 +257,12 @@ def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_le
         markdown_text += f'\n<details><summary>Click to expand Forecast Table</summary>\n\n{forecast_table}\n</details>\n\n'
         markdown_text += f'<details><summary>Click to expand Sample Historical Data</summary>\n\n{sample_data}\n</details>'
 
-        # Create plot
+        # Create plot (single subplot)
         fig, ax = plt.subplots(figsize=(14, 7))
-        ax.plot(context_df.index, context_df['sales'], label=f'Used Historical {selected_col}', color='#1f77b4', linewidth=1.5, alpha=0.8)
+        fig.set_dpi(300)  # High resolution for PNG zoom
+
+        # Historical and held-out
+        ax.plot(context_df.index, context_df['sales'], label='Historical Data', color='#1f77b4', linewidth=1.5, alpha=0.8)
         if not held_out_df.empty:
             ax.plot(held_out_df.index, held_out_df['sales'], label='Held Out Actual (Validation)', color='#2ca02c', linestyle=':', linewidth=2)
 
@@ -254,47 +279,46 @@ def run_forecast(file, time_col, selected_col, start_idx, end_idx, prediction_le
             ma = context_df['sales'].rolling(window=window, min_periods=1).mean()
             ax.plot(context_df.index, ma, label=f'Moving Average ({window} periods)', color='purple', linewidth=2)
 
-        ax.plot(pred_df.index, pred_df['predicted_sales_mean'], label='TiRex Forecast (Mean)', color='#ff7f0e', linestyle='--', linewidth=2)
+        # Median forecast: regular green line
+        ax.plot(pred_df.index, median, label='Median Forecast', color='green', linewidth=2, alpha=0.9)
 
         # Fan chart: non-overlapping bands
         # Inner 50% (lightest, center)
         ax.fill_between(pred_df.index, lower50, upper50,
-                        color='#d62728', alpha=0.1, label='50% Uncertainty')
+                         color='#d62728', alpha=0.1, label='50% Prediction Interval')
         # Wings: between 50% and slider level (medium)
         ax.fill_between(pred_df.index, lower_slider, lower50,
-                        color='#d62728', alpha=0.3)
+                         color='#d62728', alpha=0.3)
         ax.fill_between(pred_df.index, upper50, upper_slider,
-                        color='#d62728', alpha=0.3, label=f'{confidence}% Uncertainty Wings')
-
-        # Subtle skew visualization: colored segments on the median forecast line
-        from matplotlib.lines import Line2D
-        legend_elements = []
-
-        skew_colors = {'Upside': 'green', 'Downside': 'red', 'Neutral': 'gray'}
-        for i in range(len(pred_df) - 1):
-            start_date = pred_df.index[i]
-            end_date = pred_df.index[i + 1]
-            start_val = median[i]
-            end_val = median[i + 1]
-            skew = skew_directions[i]
-            color = skew_colors[skew]
-            ax.plot([start_date, end_date], [start_val, end_val], color=color, linewidth=2.5, alpha=0.7)
-
-        # Connect the last point if needed, but since segments cover, add a small marker at end if desired
-        ax.plot(pred_df.index[-1], median[-1], marker='o', color=skew_colors[skew_directions[-1]], markersize=4, alpha=0.7)
-
-        # Add to legend only if present
-        if upside_count > 0:
-            legend_elements.append(Line2D([0], [0], color='green', lw=2, label='Upside Skew'))
-        if downside_count > 0:
-            legend_elements.append(Line2D([0], [0], color='red', lw=2, label='Downside Skew'))
-        if neutral_count > 0:
-            legend_elements.append(Line2D([0], [0], color='gray', lw=2, label='Neutral Skew'))
+                         color='#d62728', alpha=0.3, label=f'{confidence}% Prediction Interval')
+
+        # Optional skew visualization on twin axis (light lines)
+        skew_handles = []
+        if add_skew_viz:
+            ax2 = ax.twinx()
+            # Light line for skew_ratio
+            line1, = ax2.plot(pred_df.index, skew_ratios, label='Skew Ratio', color='lightblue', linewidth=1, alpha=0.6)
+            skew_handles.append(line1)
+            # Light line for delta_skew (momentum) - milder color
+            line2, = ax2.plot(pred_df.index, delta_skews, label='Skew Momentum', color='lightgray', linewidth=1, alpha=0.6)
+            skew_handles.append(line2)
+            ax2.set_ylabel('Skew (-1 to 1)', color='lightblue')
+            ax2.tick_params(colors='lightblue')
+            # Set limits for visibility
+            ax2.set_ylim(-1.2, 1.2)
 
         ax.set_title(f'{selected_col} Forecast with TiRex (Context: {context_start+1}-{context_end}, Horizon: {pred_len})', fontsize=16, fontweight='bold')
         ax.set_xlabel('Date', fontsize=12)
         ax.set_ylabel(selected_col, fontsize=12)
-        ax.legend(handles=ax.get_legend_handles_labels()[0] + legend_elements, fontsize=10)
+
+        # Combined legend to avoid overlap
+        if add_skew_viz:
+            handles1, labels1 = ax.get_legend_handles_labels()
+            handles2, labels2 = ax2.get_legend_handles_labels()
+            ax.legend(handles1 + handles2, labels1 + labels2, fontsize=10, loc='upper left')
+        else:
+            ax.legend(fontsize=10)
+
         ax.tick_params(axis='x', rotation=45)
         plt.tight_layout()
 
@@ -318,7 +342,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue", secondary_hue="red"), ti
     }
     </style>
     """)
-    
+
     gr.Markdown("""
     # TiRex Forecaster Dashboard
     Upload a CSV file with a time column and numeric columns. Select the time column and one numeric column to forecast future values using the TiRex model.
@@ -361,7 +385,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue", secondary_hue="red"), ti
             )
             confidence = gr.Slider(
                 minimum=50, maximum=95, value=80, step=5,
-                label="Confidence Level (%)",
+                label="Prediction Interval (%)",
                 elem_id="confidence"
             )
             trend_checkbox = gr.Checkbox(
@@ -378,6 +402,10 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue", secondary_hue="red"), ti
                 elem_id="ma_window",
                 visible=False
             )
+            skew_checkbox = gr.Checkbox(
+                label="Add Skew Ratio & Momentum",
+                value=False
+            )
             run_button = gr.Button(
                 "Run forecast",
                 variant="primary",
@@ -414,9 +442,9 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue", secondary_hue="red"), ti
     # Event for running forecast
     run_button.click(
         run_forecast,
-        inputs=[csv_file, time_dropdown, column_dropdown, start_slider, end_slider, prediction_length, confidence, trend_checkbox, ma_checkbox, ma_slider],
+        inputs=[csv_file, time_dropdown, column_dropdown, start_slider, end_slider, prediction_length, confidence, trend_checkbox, ma_checkbox, ma_slider, skew_checkbox],
         outputs=[forecast_plot, output_text]
     )
 
-# Launch the app
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()