Week 8- Documentation

Advanced Feature Engineering and Sentiment Analysis/app.py

@@ -0,0 +1,275 @@
+import streamlit as st
+import pandas as pd
+import plotly.express as px
+import plotly.graph_objects as go
+
+# ----------------------------
+# Load Dataset
+# ----------------------------
+
+@st.cache_data
+def load_data():
+    return pd.read_csv("Football player valuation analyzer\player_transfer_value_with_sentiment.csv")
+
+df = load_data()
+
+st.title("⚽ TransferIQ Player Intelligence Dashboard")
+
+# ----------------------------
+# Player Selection
+# ----------------------------
+
+players = df["player_name"].unique()
+
+player_name = st.selectbox("Select Player", players)
+
+player_df = df[df["player_name"] == player_name]
+latest = player_df.sort_values("season").iloc[-1]
+
+# ----------------------------
+# Player Overview
+# ----------------------------
+
+st.header("Player Profile")
+
+col1, col2, col3 = st.columns(3)
+
+col1.metric("Age", latest["current_age"])
+col2.metric("Position", latest["position"])
+col3.metric("Team", latest["team"])
+
+col1.metric("Market Value (€)", int(latest["market_value_eur"]))
+col2.metric("Career Stage", latest["career_stage"])
+col3.metric("Availability Rate", round(latest["availability_rate"],2))
+
+# ----------------------------
+# Market Value Trend
+# ----------------------------
+
+st.header("Market Value Trend")
+
+fig = px.line(
+    player_df,
+    x="season",
+    y="market_value_eur",
+    markers=True,
+    title="Market Value Over Time"
+)
+
+st.plotly_chart(fig, use_container_width=True)
+
+# ----------------------------
+# Performance Stats
+# ----------------------------
+
+st.header("Performance Statistics")
+
+perf_cols = [
+    "goals",
+    "assists",
+    "shots",
+    "passes_total",
+    "passes_complete",
+    "tackles_total",
+    "interceptions",
+    "dribbles"
+]
+
+fig = px.bar(
+    x=perf_cols,
+    y=[latest[col] for col in perf_cols],
+    labels={"x":"Metric","y":"Value"},
+    title="Performance Metrics"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Per 90 Metrics
+# ----------------------------
+
+st.header("Per 90 Performance")
+
+per90_cols = [
+    "goals_per90",
+    "assists_per90",
+    "shots_per90",
+    "goal_contributions_per90",
+    "defensive_actions_per90",
+    "dribbles_per90"
+]
+
+fig = px.bar(
+    x=per90_cols,
+    y=[latest[c] for c in per90_cols],
+    title="Per 90 Contributions"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Passing Analysis
+# ----------------------------
+
+st.header("Passing Analysis")
+
+fig = px.pie(
+    values=[
+        latest["passes_complete"],
+        latest["passes_total"] - latest["passes_complete"]
+    ],
+    names=["Completed","Failed"],
+    title="Pass Completion"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Defensive Metrics
+# ----------------------------
+
+st.header("Defensive Metrics")
+
+defensive_cols = [
+    "tackles_total",
+    "tackles_won",
+    "interceptions",
+    "fouls_committed"
+]
+
+fig = px.bar(
+    x=defensive_cols,
+    y=[latest[c] for c in defensive_cols],
+    title="Defensive Contribution"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Injury Analysis
+# ----------------------------
+
+st.header("Injury History")
+
+injury_cols = [
+    "total_injuries",
+    "total_days_injured",
+    "total_matches_missed"
+]
+
+fig = px.bar(
+    x=injury_cols,
+    y=[latest[c] for c in injury_cols],
+    title="Injury Impact"
+)
+
+st.plotly_chart(fig)
+
+st.write("Most Common Injury:", latest["most_common_injury"])
+
+# ----------------------------
+# Social Media Sentiment
+# ----------------------------
+
+st.header("Social Sentiment Analysis")
+
+sentiment_cols = [
+    "positive_count",
+    "negative_count",
+    "neutral_count"
+]
+
+fig = px.pie(
+    values=[latest[c] for c in sentiment_cols],
+    names=["Positive","Negative","Neutral"],
+    title="Fan Sentiment"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Sentiment Scores
+# ----------------------------
+
+st.subheader("Sentiment Scores")
+
+fig = px.bar(
+    x=[
+        "VADER Positive",
+        "VADER Negative",
+        "VADER Compound",
+        "TextBlob Polarity",
+        "TextBlob Subjectivity"
+    ],
+    y=[
+        latest["vader_positive_score"],
+        latest["vader_negative_score"],
+        latest["vader_compound_score"],
+        latest["tb_polarity"],
+        latest["tb_subjectivity"]
+    ],
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Transfer Value Intelligence
+# ----------------------------
+
+st.header("Transfer Market Intelligence")
+
+market_cols = [
+    "social_buzz_score",
+    "transfer_attractiveness_score",
+    "tweet_engagement_rate",
+]
+
+fig = px.bar(
+    x=market_cols,
+    y=[latest[c] for c in market_cols],
+    title="Market Influence Indicators"
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Radar Chart (Overall Player Profile)
+# ----------------------------
+
+st.header("Player Radar")
+
+radar_features = [
+    "goals_per90",
+    "assists_per90",
+    "shots_per90",
+    "defensive_actions_per90",
+    "pass_accuracy_pct",
+    "dribbles_per90"
+]
+
+values = [latest[f] for f in radar_features]
+
+fig = go.Figure()
+
+fig.add_trace(go.Scatterpolar(
+    r=values,
+    theta=radar_features,
+    fill="toself",
+    name=player_name
+))
+
+fig.update_layout(
+    polar=dict(radialaxis=dict(visible=True)),
+    showlegend=False
+)
+
+st.plotly_chart(fig)
+
+# ----------------------------
+# Raw Data Table
+# ----------------------------
+
+st.header("Raw Player Data")
+
+
+st.dataframe(player_df)

Advanced Feature Engineering and Sentiment Analysis/feature_engg.py

@@ -0,0 +1,58 @@
+import pandas as pd
+import numpy as np
+
+# Load your longitudinal dataset
+df = pd.read_csv('merged_football_dataset.csv')
+
+# --- FEATURE ENGINEERING FOR MODELING ---
+
+# 1. Target Variables (What the model might predict)
+# Year-Over-Year Change in Market Value
+df['market_value_yoy_change'] = df.groupby('player_name')['market_value_eur'].diff().fillna(0)
+
+# Percentage Change in Market Value (with epsilon to prevent division by zero)
+epsilon = 1e-5
+df['market_value_yoy_pct_change'] = (
+    df['market_value_yoy_change'] / 
+    (df['market_value_eur'] - df['market_value_yoy_change'] + epsilon)
+).fillna(0).clip(lower=-1.0, upper=5.0) # Clip extreme percentages for stability
+
+# 2. Performance Metrics (Standardized to per 90 minutes)
+# Attacking Metric: Goal Involvement per 90 mins
+df['goal_involvement_per_90'] = (
+    (df['goals_total'] + df['assists_total']) / (df['minutes_played_season'] / 90)
+).fillna(0)
+
+# Replace 'inf' values that happen if a player played 0 minutes
+df.replace([np.inf, -np.inf], 0, inplace=True)
+
+# Defensive Metric: Defensive Actions per 90 mins
+df['defensive_actions_per_90'] = (
+    (df['sb_tackles'] + df['sb_interceptions']) / (df['minutes_played_season'] / 90)
+).fillna(0)
+
+# 3. Injury & Availability Metrics
+# Availability Index: 0 to 1 scale (1 means available 100% of the year)
+df['availability_index'] = (1 - (df['season_days_injured'] / 365)).clip(0, 1)
+
+# 4. Sentiment Metrics
+# Composite Sentiment Score (Average of Fan and Media)
+df['overall_sentiment'] = (df['fan_sentiment'] + df['media_sentiment']) / 2
+
+# Sentiment Trend (Is the player's reputation currently improving or worsening?)
+df['sentiment_yoy_change'] = df.groupby('player_name')['overall_sentiment'].diff().fillna(0)
+
+# --- FINAL SELECTION ---
+# Extract only the relevant features for the final modeling dataset
+final_cols = [
+    'player_name', 'season_year', 'age', 'position', 
+    'market_value_eur', 'market_value_yoy_change', 'market_value_yoy_pct_change', 
+    'minutes_played_season', 'goal_involvement_per_90', 'defensive_actions_per_90', 'sb_pass_accuracy',
+    'season_days_injured', 'injury_risk_score', 'availability_index', 
+    'fan_sentiment', 'media_sentiment', 'overall_sentiment', 'sentiment_yoy_change'
+]
+
+modeling_df = df[final_cols]
+
+# Save the final engineered features
+# modeling_df.to_csv('final_modeling_features.csv', index=False)