Update app.py

app.py

@@ -12,8 +12,6 @@
 # See the License for the specific language governing permissions and 
 # limitations under the License. 
 import spaces
-
-
 import streamlit as st
 import requests
 import pandas as pd
@@ -108,6 +106,7 @@ def page2():
         cik_input = st.text_input("Enter CIK", value="0001067983", help="Enter the Central Index Key (CIK) of the institutional investor.")
         start_date = st.date_input("Select Start Date", value=datetime(2021, 9, 30), help="Choose the start date for fetching portfolio allocation data.")
         top_n = st.number_input("Top N Groups", min_value=1, max_value=100, value=10, step=1, help="Select the number of top groups to display in trend charts.")
+        top_n_changes = st.number_input("Top N Changes", min_value=1, max_value=100, value=10, step=1, help="Select the number of top changes to display.")
     
     run_button = st.sidebar.button("Run")
 
@@ -157,6 +156,11 @@ def page2():
                 fig4 = plot_allocation_trends(allocation_data, "industryTitle", st.session_state.portfolio_allocation_params.get('top_n'))
                 st.plotly_chart(fig4, use_container_width=True, height=600)
                 
+                st.subheader("Top Changes in Portfolio Allocation")
+                fig_top_changes = plot_top_changes_plotly(allocation_data, n=top_n_changes)
+                st.plotly_chart(fig_top_changes, use_container_width=True)
+                
+                
                 # Transpose data at the bottom
                 st.markdown("---")
                 st.subheader("Transposed Ticker Data")
@@ -230,6 +234,65 @@ def page2():
         transposed_industry_data = transpose_data(allocation_data, "industryTitle")
         st.dataframe(transposed_industry_data, use_container_width=True)
 
+def plot_top_changes_plotly(df, n=10):
+    # Get unique dates in descending order
+    unique_dates = sorted(df["date"].unique(), reverse=True)
+    if len(unique_dates) < 2:
+        return go.Figure()  # Not enough data
+
+    latest_date = unique_dates[0]
+    prev_date = unique_dates[1]
+
+    # Filter data for the two most recent dates
+    df_latest = df[df["date"] == latest_date]
+    df_prev = df[df["date"] == prev_date]
+
+    # Aggregate data by symbol for each date
+    df_latest_agg = df_latest.groupby("symbol", as_index=False).agg({
+        "weight": "sum",
+        "marketValue": "sum"
+    })
+    df_prev_agg = df_prev.groupby("symbol", as_index=False).agg({
+        "weight": "sum",
+        "marketValue": "sum"
+    })
+
+    # Merge data on symbol to include new additions or removals
+    merged = pd.merge(df_prev_agg, df_latest_agg, on="symbol",
+                      suffixes=("_prev", "_latest"), how="outer")
+    merged.fillna(0, inplace=True)
+
+    # Compute changes in weight and market value
+    merged['delta_weight'] = merged['weight_latest'] - merged['weight_prev']
+    merged['delta_marketValue'] = merged['marketValue_latest'] - merged['marketValue_prev']
+
+    # Select top n positive changes and top n negative changes
+    top_up = merged[merged['delta_weight'] > 0].sort_values(
+        by='delta_weight', ascending=False).head(n)
+    top_down = merged[merged['delta_weight'] < 0].sort_values(
+        by='delta_weight', ascending=True).head(n)
+
+    combined = pd.concat([top_up, top_down])
+    combined['color'] = combined['delta_weight'].apply(lambda x: 'green' if x >= 0 else 'red')
+    combined['text'] = combined.apply(
+        lambda row: f"{row['delta_weight']:.1f}% ({format_market_value(row['delta_marketValue'])})", axis=1)
+
+    fig = go.Figure(data=[go.Bar(
+        x=combined['symbol'],
+        y=combined['delta_weight'],
+        text=combined['text'],
+        textposition='auto',
+        marker_color=combined['color']
+    )])
+    fig.update_layout(
+        title=f"Top Up and Down Changes: {prev_date} to {latest_date}",
+        xaxis_title="Symbol",
+        yaxis_title="Change in Weight (%)",
+        template="plotly_white",
+        height=500
+    )
+    return fig
+
 # Functions used in Page 2
 @st.cache_data
 def fetch_dates(cik):