Syracuse Women's Lacrosse Data Analysis - README

Overview

This repository contains a dataset (womens_lacrosse_stats.csv) and an analysis document (LLM Doc with prompts and answers.docx) for the Syracuse Women's Lacrosse team's performance in a recent season. The analysis includes 20 key questions about team performance, player statistics, and strategic insights, comparing manual answers with responses generated by an LLM (ChatGPT).

---

Files

1. DATA: womens_lacrosse_stats.csv

   • Contains player statistics, including goals (G), assists (A), points (PTS), shots (SH), shooting percentage (SH%), ground balls (GB), turnovers (TO), caused turnovers (CT), and more.
   • Key Columns:
     • PLAYER: Player name
     • GP-GS: Games played and games started
     • G, A, PTS: Goals, assists, and total points
     • SH%, SOG%: Shooting and shot-on-goal percentages
     • GWG: Game-winning goals
     • GB, TO, CT: Ground balls, turnovers, and caused turnovers

2. LLM Doc with prompts and answers.docx

   • Documents the process of querying an LLM (ChatGPT) with the dataset and refining prompts for accurate answers.
   • Includes:
     • Original questions (e.g., "Who was the leading goal scorer?").
     • Manual answers (derived from direct data analysis).
     • LLM-generated answers (with corrections where needed).
     • Notes on discrepancies and follow-up prompts.

---

Key Insights

Team Performance

• Games Played: 19
• Total Points Scored: 347 (235 goals + 112 assists)
• Top Scorer: Emma Muchnick (34 goals)
• Top Assister: Angela Beardsley (46 assists)
• Defensive Standout: Coco Vandiver (40 caused turnovers)

Efficiency & Clutch Performance

• Best Shooting % (≥20 shots): Olivia Adamson (55.6%)
• Most Game-Winning Goals: Caroline Trinkaus (4 GWG)
• Most Complete Player: Joely Caramelli (20 PTS, 23 GB, 11 CT)

Strategic Recommendations

• Recruiting Focus: Two-way midfielders (balance offense/defense).
• Development Priority: Caroline Trinkaus (improve shooting consistency).
• Team Weakness: Defense (opponents scored 337 PTS vs. Syracuse’s 347).

---

How to Use This Data

• Identify top performers and areas for improvement (e.g., shooting efficiency, turnovers).
• Compare player contributions (e.g., PTS/GP for starters vs. bench).
• Simulate scenarios (e.g., impact of a 10% shooting percentage increase).
• Study correlations (e.g., shots vs. goals: r = 0.99).
• Test LLM accuracy in sports analytics by replicating prompts.
• Practice data cleaning (e.g., handling duplicate entries like "Peers, Ava").
• Explore statistical metrics (e.g., GB/TO ratio for possession analysis).

---

Notes

• Data Limitations:
  • No game-by-game trends (aggregate stats only).
  • Small sample sizes for some players (e.g., Olivia Adamson: 3 GP).
• LLM Challenges:
  • Required follow-up prompts for accuracy (e.g., "Double-check assists").
  • Misinterpreted player names (e.g., "Ward, Emma" vs. "Beardsley, Angela").