AthletiQ - Basketball Training App

What it does

AthletiQ is a mobile basketball training app that turns solo practice into an exciting game simulation. Players create profiles with team selection, position, and difficulty level, then play 20-shot "games" against AI opponents with realistic make rates.

Key Features

Game atmosphere: Real-time scoreboard, quarter system, and shot clock
Audio prompts: Voice announcements for each shot type
Smart shot selection: Database of 20+ basketball shots that adapt to game situation
AI shot detection: Computer vision that automatically detects makes/misses (optional)
Voice recognition: Players can say "make" or "miss" instead of pressing buttons
Stats tracking: Complete season statistics with shooting percentages and game history
Multiple difficulty levels: From Rookie (35% opponent make rate) to MVP (75% make rate)

How we built it

Technology Stack

Frontend: React Native with Expo for cross-platform mobile development
Backend: Supabase for authentication, database, and real-time features
AI Vision: Python backend with YOLOv8 model for basketball/hoop detection
Audio: Text-to-speech for shot prompts and pre-recorded audio clips
Database: PostgreSQL with tables for users, games, shots, teams, and statistics

Key Technical Components

Game engine that simulates realistic basketball scoring and possession flow
WebSocket connection between mobile app and Python AI server
Real-time camera processing for automatic shot detection
Voice recognition using OpenAI Whisper API
Dynamic shot selection based on game situation and difficulty

Challenges we ran into

AI Integration

Getting the computer vision model to work reliably with mobile cameras was difficult. We had to optimize frame processing and handle different lighting conditions.

Real-time Synchronization

Coordinating between the mobile app, camera feed, AI processing, and database updates while maintaining smooth gameplay was complex.

Audio System

Implementing both text-to-speech and pre-recorded audio clips while managing timing with game flow and shot detection.

Game Balance

Making the AI opponents feel realistic at different difficulty levels required extensive testing and tweaking of make rates and possession simulation.

Cross-platform Compatibility

Ensuring the app works smoothly on both iOS and Android with camera permissions and audio features.

Accomplishments that we're proud of

Complete MVP delivered: Full game experience from login to stats tracking works perfectly
Innovative AI integration: Successfully combined computer vision with mobile gaming for automatic shot detection
Realistic game simulation: The opponent AI and game flow feel authentic to actual basketball
Seamless user experience: Players can use voice commands, manual buttons, or AI detection interchangeably
Professional polish: Quality audio, smooth animations, and NBA-style presentation
Scalable architecture: Clean code structure that can easily add multiplayer, more teams, and advanced features

What we learned

Mobile-AI Integration

How to efficiently stream camera data to an AI model and get real-time results back to the mobile app.

Game Design Psychology

Balancing difficulty and engagement to keep players motivated without making it too easy or frustrating.

Real-time Systems

Managing multiple concurrent processes (camera, audio, AI, database) while maintaining app performance.

User Experience Design

Creating intuitive interfaces for sports applications where quick actions and clear feedback are essential.

Database Optimization

Structuring data for both real-time gameplay and historical analytics without performance issues.

What's next for AthletiQ

Short-term Goals

Multiplayer mode for friends to compete remotely
More teams and customization options
Advanced shot charts and analytics
Social features like leaderboards and sharing

Long-term Vision

Professional training programs with skill progression
Integration with wearable devices for movement analysis
AR features for virtual coaching
Partnership with basketball organizations for official content
Machine learning to personalize difficulty and shot selection based on player improvement