Update app.py

app.py

@@ -354,78 +354,159 @@ class TyphoonAnalyzer:
         }
 
     def create_tracks_plot(self, data):
-        """Create typhoon tracks visualization"""
-        fig = go.Figure()
-        
-        fig.update_layout(
-            title={
-                'text': 'Typhoon Tracks',
-                'y':0.95,
-                'x':0.5,
-                'xanchor': 'center',
-                'yanchor': 'top'
-            },
-            showlegend=True,
-            legend=dict(
-                yanchor="top",
-                y=0.99,
-                xanchor="left",
-                x=0.01,
-                bgcolor='rgba(255, 255, 255, 0.8)'
-            ),
-            geo=dict(
-                projection_type='mercator',
-                showland=True,
-                showcoastlines=True,
-                landcolor='rgb(243, 243, 243)',
-                countrycolor='rgb(204, 204, 204)',
-                coastlinecolor='rgb(214, 214, 214)',
-                showocean=True,
-                oceancolor='rgb(230, 250, 255)',
-                showlakes=True,
-                lakecolor='rgb(230, 250, 255)',
-                lataxis=dict(range=[0, 50]),
-                lonaxis=dict(range=[100, 180]),
-                center=dict(lat=20, lon=140),
-                bgcolor='rgba(255, 255, 255, 0.5)'
-            ),
-            paper_bgcolor='rgba(255, 255, 255, 0.5)',
-            plot_bgcolor='rgba(255, 255, 255, 0.5)'
-        )
-
-        for category in COLOR_MAP.keys():
-            category_data = data[data['Category'] == category]
-            
-            # Group by SID to get individual typhoon tracks
-            for _, storm in category_data.groupby('SID'):
-                # Create single track for each typhoon
-                track_data = self.typhoon_data[self.typhoon_data['SID'] == storm['SID'].iloc[0]]
-                track_data = track_data.sort_values('ISO_TIME')
-                
-                fig.add_trace(go.Scattergeo(
-                    lon=track_data['LON'],
-                    lat=track_data['LAT'],
-                    mode='lines',
-                    line=dict(
-                        width=2,
-                        color=COLOR_MAP[category]
-                    ),
-                    name=category,
-                    legendgroup=category,
-                    showlegend=True if storm.iloc[0]['SID'] == category_data.iloc[0]['SID'] else False,
-                    hovertemplate=(
-                        f"Name: {storm['NAME'].iloc[0]}<br>" +
-                        f"Category: {category}<br>" +
-                        f"Wind Speed: {storm['USA_WIND'].iloc[0]:.1f} kt<br>" +
-                        f"Pressure: {storm['WMO_PRES'].iloc[0]:.1f} hPa<br>" +
-                        f"Date: {track_data['ISO_TIME'].dt.strftime('%Y-%m-%d %H:%M').iloc[0]}<br>" +
-                        f"Lat: %{lat:.2f}°N<br>" +
-                        f"Lon: %{lon:.2f}°E<br>" +
-                        "<extra></extra>"
-                    )
-                ))
-
-        return fig
+       """Create typhoon tracks visualization"""
+       fig = go.Figure()
+       
+       fig.update_layout(
+           title={
+               'text': 'Typhoon Tracks',
+               'y':0.95,
+               'x':0.5,
+               'xanchor': 'center',
+               'yanchor': 'top'
+           },
+           showlegend=True,
+           legend=dict(
+               yanchor="top",
+               y=0.99,
+               xanchor="left",
+               x=0.01,
+               bgcolor='rgba(255, 255, 255, 0.8)'
+           ),
+           geo=dict(
+               projection_type='mercator',
+               showland=True,
+               showcoastlines=True,
+               landcolor='rgb(243, 243, 243)',
+               countrycolor='rgb(204, 204, 204)',
+               coastlinecolor='rgb(214, 214, 214)',
+               showocean=True,
+               oceancolor='rgb(230, 250, 255)',
+               showlakes=True,
+               lakecolor='rgb(230, 250, 255)',
+               lataxis=dict(range=[0, 50]),
+               lonaxis=dict(range=[100, 180]),
+               center=dict(lat=20, lon=140),
+               bgcolor='rgba(255, 255, 255, 0.5)'
+           ),
+           paper_bgcolor='rgba(255, 255, 255, 0.5)',
+           plot_bgcolor='rgba(255, 255, 255, 0.5)'
+       )
+
+       for category in COLOR_MAP.keys():
+           category_data = data[data['Category'] == category]
+           for _, storm in category_data.groupby('SID'):
+               track_data = self.typhoon_data[self.typhoon_data['SID'] == storm['SID'].iloc[0]]
+               track_data = track_data.sort_values('ISO_TIME')
+               
+               fig.add_trace(go.Scattergeo(
+                   lon=track_data['LON'],
+                   lat=track_data['LAT'],
+                   mode='lines',
+                   line=dict(
+                       width=2,
+                       color=COLOR_MAP[category]
+                   ),
+                   name=category,
+                   legendgroup=category,
+                   showlegend=True if storm.iloc[0]['SID'] == category_data.iloc[0]['SID'] else False,
+                   hovertemplate=(
+                       f"Name: {storm['NAME'].iloc[0]}<br>" +
+                       f"Category: {category}<br>" +
+                       f"Wind Speed: {storm['USA_WIND'].iloc[0]:.1f} kt<br>" +
+                       f"Pressure: {storm['WMO_PRES'].iloc[0]:.1f} hPa<br>" +
+                       f"Date: {track_data['ISO_TIME'].dt.strftime('%Y-%m-%d %H:%M').iloc[0]}<br>" +
+                       f"Lat: {track_data['LAT'].iloc[0]:.2f}°N<br>" +
+                       f"Lon: {track_data['LON'].iloc[0]:.2f}°E<br>" +
+                       "<extra></extra>"
+                   )
+               ))
+
+       return fig
+
+   def analyze_clusters(self, year, n_clusters):
+       """Analyze typhoon clusters for a specific year"""
+       year_data = self.typhoon_data[self.typhoon_data['SEASON'] == year]
+       if year_data.empty:
+           return go.Figure(), "No data available for selected year"
+
+       # Prepare data for clustering
+       routes = []
+       for _, storm in year_data.groupby('SID'):
+           if len(storm) > 1:
+               # Standardize route length
+               t = np.linspace(0, 1, len(storm))
+               t_new = np.linspace(0, 1, 100)
+               lon_interp = interp1d(t, storm['LON'], kind='linear')(t_new)
+               lat_interp = interp1d(t, storm['LAT'], kind='linear')(t_new)
+               routes.append(np.column_stack((lon_interp, lat_interp)))
+       
+       if len(routes) < n_clusters:
+           return go.Figure(), f"Not enough typhoons ({len(routes)}) for {n_clusters} clusters"
+
+       # Perform clustering
+       routes_array = np.array(routes)
+       routes_reshaped = routes_array.reshape(routes_array.shape[0], -1)
+       kmeans = KMeans(n_clusters=n_clusters, random_state=42)
+       clusters = kmeans.fit_predict(routes_reshaped)
+       
+       # Create visualization
+       fig = go.Figure()
+       
+       # Set layout
+       fig.update_layout(
+           title=f'Typhoon Route Clusters ({year})',
+           showlegend=True,
+           geo=dict(
+               projection_type='mercator',
+               showland=True,
+               showcoastlines=True,
+               landcolor='rgb(243, 243, 243)',
+               countrycolor='rgb(204, 204, 204)',
+               coastlinecolor='rgb(214, 214, 214)',
+               showocean=True,
+               oceancolor='rgb(230, 250, 255)',
+               lataxis=dict(range=[0, 50]),
+               lonaxis=dict(range=[100, 180]),
+               center=dict(lat=20, lon=140)
+           )
+       )
+       
+       # Plot routes colored by cluster
+       for route, cluster_id in zip(routes, clusters):
+           fig.add_trace(go.Scattergeo(
+               lon=route[:, 0],
+               lat=route[:, 1],
+               mode='lines',
+               line=dict(
+                   width=1, 
+                   color=f'hsl({cluster_id * 360/n_clusters}, 50%, 50%)'
+               ),
+               name=f'Cluster {cluster_id + 1}',
+               showlegend=False
+           ))
+       
+       # Plot cluster centers
+       for i in range(n_clusters):
+           center = kmeans.cluster_centers_[i].reshape(-1, 2)
+           fig.add_trace(go.Scattergeo(
+               lon=center[:, 0],
+               lat=center[:, 1],
+               mode='lines',
+               name=f'Cluster {i+1} Center',
+               line=dict(
+                   width=3,
+                   color=f'hsl({i * 360/n_clusters}, 100%, 50%)'
+               )
+           ))
+       
+       # Generate statistics text
+       stats_text = "### Clustering Results\n\n"
+       cluster_counts = np.bincount(clusters)
+       for i in range(n_clusters):
+           stats_text += f"- Cluster {i+1}: {cluster_counts[i]} typhoons\n"
+       
+       return fig, stats_text
     def get_typhoons_for_year(self, year):
         """Get list of typhoons for a specific year"""
         year_data = self.typhoon_data[self.typhoon_data['ISO_TIME'].dt.year == year]
@@ -830,7 +911,7 @@ def create_interface():
 
         def update_typhoon_choices(year):
             typhoons = analyzer.get_typhoons_for_year(year)
-            return gr.Dropdown.update(choices=typhoons, value=None)
+            return gr.update(choices=typhoons, value=None)
 
         # Connect events for main analysis
         analyze_btn.click(