Counter-Strike: Global Offensive Champions Match Tracker

A MongoDB-based data science project that tracks and analyses CS:GO championship matches using web scraping and aggregation pipelines.

---

Project Overview

During the COVID-19 lockdown, Counter-Strike: Global Offensive (CS:GO) served as more than a game — it was a way to connect with friends. Years later in university, inspired by those shared experiences and a memorable esports tournament (PGL Major Antwerp 2022), this project explores CS:GO match data by sourcing, storing, and analysing statistics in a structured MongoDB environment.

---

Dataset Description

• Source: Scraped from bo3.gg
• Matches Analysed: 2022 PGL Major matches
• Data Type: Match metadata, teams, players, maps, stats, and simulated commentary

---

Schema Design

The project uses six MongoDB collections:

matches

• match_id, title, tournament, date
• teams: list of team IDs
• score, map_ids

teams

• team_id, team_name
• players: list of player IDs
• tournament

players

• player_id, name, tournament, team_id

maps

• map_id, name, times_played

player_stats

• match_id, player_id, kills, deaths, assists

commentary

• commentary_id, match_id, timestamp, text, tags

All keys are uniquely generated using UUIDs for consistency and ease of reference.

---

Data Ingestion

• Web scraper built using:
  • BeautifulSoup4, requests, uuid, pymongo
• Robust error-handling and duplicate prevention mechanisms
• HTML parsing logic refined over 30+ iterations

---

Live Commentary Simulation

Since timestamped highlights were unavailable, a 2-minute mock simulation of live commentary was scripted. See the .ipynb notebook for execution instructions.

---

Data Verification Functions

Implemented in Python with pymongo:

• verify_single_team_per_player()
• verify_teams_in_matches()
• verify_player_stats_references()
• verify_maps_in_matches()
• verify_commentary_matches()

These ensure referential integrity across the dataset.

---

Aggregation Queries

1. Full Match Summary
   Nested lookups and array queries to compile match data, participating teams, players, and commentary.

2. Top KDA Players
   Identifies top players by computing KDA = (kills + assists) / deaths.

3. Top Performer per Team/Match
   Lists the highest KDA scorer from each team in every match.

Additional explanation is available in the Jupyter Notebook.

---

Project Files

• match_tracker.ipynb: Code for scraping, ingestion, and aggregation
• README.md: This file
• data/: (optional) Any manually backed up datasets

---

References

• bo3.gg: FaZe vs Natus Vincere (22 May 2022)
• bo3.gg: Spirit vs Furia (19 May 2022)
• bo3.gg: ENCE vs Natus Vincere (21 May 2022)
• bo3.gg: FaZe vs Spirit (21 May 2022)