ScoutEdge

Turn GRID match data into instant scouting reports — player profiles, team strategies, weaknesses, and counter-tactics in under 60 seconds.

Comment

Inspiration

Before every match, esports coaches spend hours watching VODs, taking notes, and manually building scouting reports. I wanted to automate this — give coaches a tool that ingests real match data and outputs actionable intelligence in seconds, not hours.

What it does

ScoutEdge connects to GRID's Historical Data API and generates comprehensive scouting reports for any Valorant team. In under 60 seconds, you get:

  • Player Profiles: K/D ratios, agent pools, and playstyle tendencies
  • Team Strategy: Win rates, map pool analysis, and attack/defense patterns
  • Compositions: Preferred agent setups extracted from match history
  • Critical Weaknesses: Data-backed exploitable patterns (over-extension, slow rotations)
  • How to Win: Per-player counter-strategies with specific trigger conditions

Every insight includes sample size and links back to real GRID Series IDs for verification.

How I built it

  • Frontend: Next.js 15 with TypeScript and Tailwind CSS for a premium, responsive dashboard
  • Data Layer: Direct integration with GRID's GraphQL API to fetch real match data
  • Stats Engine: Custom logic to calculate K/D, win rates, agent frequency, and role-based tendencies
  • Dual Mode: Demo mode with fixture data for offline testing + live GRID mode for real data
  • Development: JetBrains WebStorm with Junie AI for code generation and refactoring

Challenges I ran into

  1. GRID API Schema: Understanding the nested structure of series → games → players took iteration
  2. Data Consistency: Ensuring insights were varied and role-specific, not generic templates
  3. Evidence Transparency: Every claim needed traceable proof — sample sizes and Series IDs
  4. Time Pressure: Balancing polish with functionality under hackathon deadlines

Accomplishments I'm proud of

  • Real GRID data integration with Series ID 2629392 from VCT Americas Kickoff
  • 15 unique player tendency combinations (K/D × Role)
  • Per-player counter-strategies that are actually actionable
  • Clean, judge-proof UI with evidence verification

What I learned

  • How to work with GRID's esports data infrastructure
  • The importance of making AI-adjacent tools transparent about their data sources
  • Building "confidence scores" based on sample size, not arbitrary percentages

What's next for ScoutEdge

  • Live match integration for real-time scouting during drafts
  • League of Legends support using GRID's LoL data
  • Team comparison mode (head-to-head analysis)
  • Export to coaching platforms and team management tools

Built With

Share this project:

Updates