Update app.py

app.py

@@ -2,8 +2,16 @@ import streamlit as st
 import pandas as pd
 import joblib
 from sklearn.ensemble import RandomForestRegressor
-import plotly.express as px
 import plotly.graph_objects as go
+from PIL import Image
+import plotly.express as px
+
+# Set the page configuration
+st.set_page_config(
+    page_title="NBA Player Performance Predictor",
+    page_icon="🏀",
+    layout="centered"
+)
 
 # Mapping for position to numeric values
 position_mapping = {
@@ -14,7 +22,7 @@ position_mapping = {
     "C": 5.0,
 }
 
-# Predefined injury types
+# Injury types and average days
 injury_types = [
     "foot fracture injury", "hip flexor surgery injury", "calf strain injury",
     "quad injury injury", "shoulder sprain injury", "foot sprain injury",
@@ -28,10 +36,66 @@ injury_types = [
 
 average_days_injured = {
     "foot fracture injury": 207.666667,
-    # ... (other injuries with their averages)
+    "hip flexor surgery injury": 256.000000,
+    "calf strain injury": 236.000000,
+    "quad injury injury": 283.000000,
+    "shoulder sprain injury": 259.500000,
+    "foot sprain injury": 294.000000,
+    "torn rotator cuff injury injury": 251.500000,
+    "torn mcl injury": 271.000000,
+    "hip flexor strain injury": 253.000000,
+    "fractured leg injury": 250.250000,
+    "sprained mcl injury": 228.666667,
+    "ankle sprain injury": 231.333333,
+    "hamstring injury injury": 220.000000,
+    "meniscus tear injury": 201.250000,
+    "torn hamstring injury": 187.666667,
+    "dislocated shoulder injury": 269.000000,
+    "ankle fracture injury": 114.500000,
+    "fractured hand injury": 169.142857,
+    "bone spurs injury": 151.500000,
+    "acl tear injury": 268.000000,
+    "hip labrum injury": 247.500000,
+    "back surgery injury": 215.800000,
+    "arm injury injury": 303.666667,
+    "torn shoulder labrum injury": 195.666667,
     "lower back spasm injury": 234.000000,
 }
 
+team_logo_paths = {
+    "Cleveland Cavaliers": "NBA_LOGOs/Clevelan-Cavaliers-logo-2022.png",
+    "Atlanta Hawks": "NBA_LOGOs/nba-atlanta-hawks-logo.png",
+    "Boston Celtics": "NBA_LOGOs/nba-boston-celtics-logo.png",
+    "Brooklyn Nets": "NBA_LOGOs/nba-brooklyn-nets-logo.png",
+    "Charlotte Hornets": "NBA_LOGOs/nba-charlotte-hornets-logo.png",
+    "Chicago Bulls": "NBA_LOGOs/nba-chicago-bulls-logo.png",
+    "Dallas Mavericks": "NBA_LOGOs/nba-dallas-mavericks-logo.png",
+    "Denver Nuggets": "NBA_LOGOs/nba-denver-nuggets-logo-2018.png",
+    "Detroit Pistons": "NBA_LOGOs/nba-detroit-pistons-logo.png",
+    "Golden State Warriors": "NBA_LOGOs/nba-golden-state-warriors-logo-2020.png",
+    "Houston Rockets": "NBA_LOGOs/nba-houston-rockets-logo-2020.png",
+    "Indiana Pacers": "NBA_LOGOs/nba-indiana-pacers-logo.png",
+    "LA Clippers": "NBA_LOGOs/nba-la-clippers-logo.png",
+    "Los Angeles Lakers": "NBA_LOGOs/nba-los-angeles-lakers-logo.png",
+    "Memphis Grizzlies": "NBA_LOGOs/nba-memphis-grizzlies-logo.png",
+    "Miami Heat": "NBA_LOGOs/nba-miami-heat-logo.png",
+    "Milwaukee Bucks": "NBA_LOGOs/nba-milwaukee-bucks-logo.png",
+    "Minnesota Timberwolves": "NBA_LOGOs/nba-minnesota-timberwolves-logo.png",
+    "New Orleans Pelicans": "NBA_LOGOs/nba-new-orleans-pelicans-logo.png",
+    "New York Knicks": "NBA_LOGOs/nba-new-york-knicks-logo.png",
+    "Oklahoma City Thunder": "NBA_LOGOs/nba-oklahoma-city-thunder-logo.png",
+    "Orlando Magic": "NBA_LOGOs/nba-orlando-magic-logo.png",
+    "Philadelphia 76ers": "NBA_LOGOs/nba-philadelphia-76ers-logo.png",
+    "Phoenix Suns": "NBA_LOGOs/nba-phoenix-suns-logo.png",
+    "Portland Trail Blazers": "NBA_LOGOs/nba-portland-trail-blazers-logo.png",
+    "Sacramento Kings": "NBA_LOGOs/nba-sacramento-kings-logo.png",
+    "San Antonio Spurs": "NBA_LOGOs/nba-san-antonio-spurs-logo.png",
+    "Toronto Raptors": "NBA_LOGOs/nba-toronto-raptors-logo-2020.png",
+    "Utah Jazz": "NBA_LOGOs/nba-utah-jazz-logo.png",
+    "Washington Wizards": "NBA_LOGOs/nba-washington-wizards-logo.png",
+}
+
+# Caching player data and model
 @st.cache_resource
 def load_player_data():
     return pd.read_csv("player_data.csv")
@@ -40,83 +104,151 @@ def load_player_data():
 def load_rf_model():
     return joblib.load("rf_injury_change_model.pkl")
 
-# Main Streamlit app
+# Main application
+# Main application
 def main():
     st.title("NBA Player Performance Predictor 🏀")
-    st.write("""
+    st.markdown("""
     Use this tool to predict how a player's performance metrics might change
     if they experience a hypothetical injury.
     """)
 
+    # Load data and model
     player_data = load_player_data()
     rf_model = load_rf_model()
 
-    # Layout: Sidebar for Inputs
+    # Sidebar inputs
     with st.sidebar:
         st.header("Player & Injury Inputs")
         player_list = sorted(player_data['player_name'].dropna().unique())
         player_name = st.selectbox("Select Player", player_list)
 
-        # Player Details
         if player_name:
+            # Filter data for the selected player
             player_row = player_data[player_data['player_name'] == player_name]
+            team_name = player_row.iloc[0]['team_abbreviation']
             position = player_row.iloc[0]['position']
-            position_numeric = position_mapping.get(position, 0)
-            st.write(f"**Position**: {position} (Numeric: {position_numeric})")
-            
             stats_columns = ['age', 'player_height', 'player_weight']
-            default_stats = {stat: player_row.iloc[0][stat] for stat in stats_columns}
 
+            st.write(f"**Position**: {position}")
+            st.write(f"**Team**: {team_name}")
+
+            # Player stats
+            default_stats = {stat: player_row.iloc[0][stat] for stat in stats_columns}
             for stat in default_stats.keys():
                 default_stats[stat] = st.number_input(f"{stat}", value=default_stats[stat])
 
+            # Injury details
             injury_type = st.selectbox("Select Hypothetical Injury", injury_types)
             default_days_injured = average_days_injured.get(injury_type, 30)
             days_injured = st.slider("Estimated Days Injured", 0, 365, int(default_days_injured))
             injury_occurrences = st.number_input("Injury Occurrences", min_value=0, value=1)
 
+            # Prepare data for prediction
             input_data = pd.DataFrame([{
                 "days_injured": days_injured,
                 "injury_occurrences": injury_occurrences,
-                "position": position_numeric,
+                "position": position_mapping.get(position, 0),
                 "injury_type": injury_type,
                 **default_stats
             }])
             input_data["injury_type"] = pd.factorize(input_data["injury_type"])[0]
 
-    # Main Layout
-    col1, col2 = st.columns(2)
+            st.header("Prediction Results")
+            if st.button("Predict"):
+                predictions = rf_model.predict(input_data)
+                predictions = [round(float(pred), 2) for pred in predictions]
+
+                # Display prediction results
+                prediction_columns = ["Predicted Change in PTS", "Predicted Change in REB", "Predicted Change in AST"]
+                result_df = pd.DataFrame([predictions], columns=prediction_columns)
+                st.table(result_df)
+            
+                # Main content layout
+    st.divider()
+    st.header("Player Overview")
+    col1, col2 = st.columns([1, 2])
 
     with col1:
         st.subheader("Player Details")
         st.metric("Age", default_stats['age'])
         st.metric("Height (cm)", default_stats['player_height'])
         st.metric("Weight (kg)", default_stats['player_weight'])
-        st.image("player_placeholder.png", caption=f"{player_name}", width=200)  # Placeholder image
 
     with col2:
-        st.subheader("Prediction Results")
-
-        if st.button("Predict"):
-            predictions = rf_model.predict(input_data)
-            predictions = [round(float(pred), 2) for pred in predictions]
-
-            prediction_columns = ["Predicted Change in PTS", "Predicted Change in REB", "Predicted Change in AST"]
-            result_df = pd.DataFrame([predictions], columns=prediction_columns)
-            
-            # Bar Chart for Visualizing Predictions
-            fig = go.Figure(data=[
-                go.Bar(
-                    x=prediction_columns,
-                    y=predictions,
-                    marker_color=["green" if val > 0 else "red" for val in predictions]
-                )
-            ])
-            fig.update_layout(title="Predicted Performance Changes", xaxis_title="Metrics", yaxis_title="Change")
-            st.plotly_chart(fig)
-
-            # Display predictions as a table
-            st.table(result_df)
+        # Display team logo
+        if team_name in team_logo_paths:
+            logo_path = team_logo_paths[team_name]
+            try:
+                logo_image = Image.open(logo_path)
+                st.image(logo_image, caption=f"{team_name} Logo", use_column_width=True)
+            except FileNotFoundError:
+                st.error(f"Logo for {team_name} not found.")
+
+
+    # Graphs for PPG, AST, and REB
+    st.divider()
+    st.header("Player Performance Graphs")
+
+    if st.button("Show Performance Graphs"):
+        # Filter data for the selected player
+        player_data_filtered = player_data[player_data["player_name"] == player_name].sort_values(by="season")
+
+        # Ensure all seasons are included
+        all_seasons = pd.Series(range(player_data["season"].min(), player_data["season"].max() + 1))
+        player_data_filtered = (
+            pd.DataFrame({"season": all_seasons})
+            .merge(player_data_filtered, on="season", how="left")
+            .fillna({"pts": 0, "ast": 0, "reb": 0})  # Fill missing values
+        )
+
+        if not player_data_filtered.empty:
+            # PPG Graph
+            fig_ppg = px.line(
+                player_data_filtered,
+                x="season",
+                y="pts",
+                title=f"{player_name}: Points Per Game (PPG) Over Seasons",
+                labels={"pts": "Points Per Game (PPG)", "season": "Season"},
+                markers=True
+            )
+            fig_ppg.update_layout(template="plotly_white")
+
+            # AST Graph
+            fig_ast = px.line(
+                player_data_filtered,
+                x="season",
+                y="ast",
+                title=f"{player_name}: Assists Per Game (AST) Over Seasons",
+                labels={"ast": "Assists Per Game (AST)", "season": "Season"},
+                markers=True
+            )
+            fig_ast.update_layout(template="plotly_white")
+
+            # REB Graph
+            fig_reb = px.line(
+                player_data_filtered,
+                x="season",
+                y="reb",
+                title=f"{player_name}: Rebounds Per Game (REB) Over Seasons",
+                labels={"reb": "Rebounds Per Game (REB)", "season": "Season"},
+                markers=True
+            )
+            fig_reb.update_layout(template="plotly_white")
+
+            # Display graphs
+            st.plotly_chart(fig_ppg, use_container_width=True)
+            st.plotly_chart(fig_ast, use_container_width=True)
+            st.plotly_chart(fig_reb, use_container_width=True)
+        else:
+            st.error("No data available for the selected player.")
+
+    # Footer
+    st.divider()
+    st.markdown("""
+    ### About This Tool
+    This application predicts how injuries might impact an NBA player's performance using machine learning models. Data is based on historical player stats and injuries.
+    """)
 
 if __name__ == "__main__":
     main()