Add FPL dashboard with player analytics and transfer insights

[328345]

Summary

This PR introduces a comprehensive Fantasy Premier League (FPL) analytics dashboard built with Streamlit, featuring player rankings, transfer activity analysis, fixture difficulty ratings, and performance trend tracking.

Key Changes

Core Data Loading (dashboard/data_loader.py)

• Implemented cached data loaders for players, teams, player stats, and gameweek summaries with 30-minute TTL
• Created get_master_player_df() to join player stats with team and player metadata
• Added get_upcoming_fixtures() to retrieve upcoming matches with Elo-based difficulty ratings
• Implemented get_recent_form_data() to load per-gameweek performance history for trend analysis

Dashboard Views

• Player Rankings (player_rankings.py): Sortable player table with multiple metrics (form, PPG, xG, xA, value, etc.) and search functionality
• Fixture Difficulty (fixture_difficulty.py): Color-coded fixture grid using Elo ratings to visualize match difficulty across upcoming gameweeks
• Value Picks (value_picks.py): Position-based value analysis with points-per-million metrics and price vs. points scatter plot
• Form Trends (form_trends.py): Multi-player performance tracking with gameweek-by-gameweek points and expected stats visualization
• Player Comparison (player_comparison.py): Side-by-side stats comparison with normalized radar charts for up to 3 players
• Transfer Activity (transfer_activity.py): Transfer in/out rankings, price changes (gameweek and season), and gameweek highlights

Main Application (dashboard/app.py)

• Streamlit app with sidebar navigation across 6 analytics pages
• Global filters for position, team, price range, and minimum minutes played
• Responsive layout with wide page configuration

Data Synchronization (sync_to_firestore.py)

• Automated script to sync CSV data to Firestore
• Discovers latest season folder and processes all gameweek directories
• Batch uploads with 499-operation chunking to respect Firestore limits
• Clears existing subcollections before syncing new data

CI/CD (github/workflows/sync.yml)

• GitHub Actions workflow for automated Firestore synchronization
• Triggers on push to main branch or manual dispatch
• Uses Workload Identity Federation for secure GCP authentication

Notable Implementation Details

• All data loaders use Streamlit's @st.cache_data decorator with 30-minute TTL for performance
• Elo-based fixture difficulty uses color gradients (dark green for easy → dark red for hard)
• Radar chart normalization uses min/max scaling across the full dataset (0-100 range)
• Position names shortened to standard FPL abbreviations (GKP, DEF, MID, FWD)
• Supports double gameweeks with multiple fixtures per team per gameweek
• Firestore sync handles null values and uses auto-generated document IDs

https://claude.ai/code/session_01FiVSPeDMHuDNYTzLuFxC26