• 

  Software Architecture

• 

  Flow Chart

• 

  Sequence Diagram

• 

• 

• 

• 

📌 Comprehensive Project Compilation

🎯 Project Objective

The aim of this project is to revolutionize the MLB fan experience through a personalized mobile application that generates and delivers real-time game summaries in multiple formats:

• ✅ Highlight video clips 📽️
• ✅ Automatically generated audio narrations 🎤
• ✅ Translations in English, Spanish, and Japanese 🌍

This allows fans to receive relevant content tailored to their preferences without the need to watch the game live, utilizing a system based on artificial intelligence and cloud computing.

🏗️ Project Overview

This system is built upon:

• Real-time data extraction and analysis from the MLB's GUMBO API.
• Processing and filtering of key events such as home runs, strikeouts, and decisive plays.
• Automated multimedia content generation using Google Cloud AI (Video Intelligence, Text-to-Speech, Translation API).
• Scalable storage and distribution via Google Cloud Storage and Cloud CDN.
• Personalized user experience, providing summaries based on team and player preferences.

🌿 Repository Branches

Our repository is organized into the following branches:

• main: Contains the stable version of the project.
• backend: Focuses on the server-side components, including:
  • Node.js + Express.js → RESTful backend server.
  • Cloud Spanner → Relational database for storing users, preferences, and games.
  • Firestore → NoSQL database for storing events and summaries.
  • Redis (Memorystore) → Cache to improve performance and reduce latency.
  • Google Tasks + Scheduler → Automation of real-time data processing.
  • Cloud Pub/Sub + Eventarc → Notification of key events in games.
  • Google Cloud Logging & Monitoring → Infrastructure supervision and diagnostics.
• frontend: Dedicated to the mobile application, featuring:
  • React Native → Main framework for mobile development.
  • Expo → Enables rapid testing and agile deployments without manual compilation.
  • Axios → HTTP client for backend communication.
  • Redux → Global state management of the app.
  • EAS (Expo Application Services) → For generating demo builds and OTA distribution.
• testing: Contains scripts to test various APIs and their response structures, primarily focusing on the MLB's GUMBO API.

⚙️ Technologies Used

📌 Backend (Express.js + Google Cloud)

• Node.js + Express.js → RESTful backend server.
• Cloud Spanner → Relational database for storing users, preferences, and games.
• Firestore → NoSQL database for storing events and summaries.
• Redis (Memorystore) → Cache to improve performance and reduce latency.
• Google Tasks + Scheduler → Automation of real-time data processing.
• Cloud Pub/Sub + Eventarc → Notification of key events in games.
• Google Cloud Logging & Monitoring → Infrastructure supervision and diagnostics.

📌 Frontend (React Native + Expo)

• React Native → Main framework for mobile development.
• Expo → Enables rapid testing and agile deployments without manual compilation.
• Axios → HTTP client for backend communication.
• Redux → Global state management of the app.
• EAS (Expo Application Services) → For generating demo builds and OTA distribution.

📌 Artificial Intelligence (Google Cloud AI)

• Vertex AI → Trains Machine Learning models to personalize summaries based on user history.
• Google Video Intelligence API → Automatically trims clips of key moments in games.
• Google Text-to-Speech API → Generates automatic narrations in English, Spanish, and Japanese.
• Google Translation API → Translates content to offer a multilingual experience.

🔄 Complete System Workflow

1. User Configuration

   • The user registers in the app.
   • Sets preferences: teams, players, types of plays.
   • Google Tasks queues a task to associate preferences with MLB data.
   • Cloud Spanner stores the user's preferences.

2. MLB Data Capture

   • Google Cloud Scheduler queries the GUMBO API every 10 minutes for scheduled games and every minute for live games.
   • Game and event data are analyzed and stored in Firestore.

3. Key Event Processing

   • Cloud Pub/Sub detects relevant events in Firestore (gameEvents).
   • Multimedia content generation processes are triggered:
     • 📽️ Clip trimming with Video Intelligence API
     • 🎤 Audio narrations with Text-to-Speech API
     • 🌍 Translation with Translation API

4. Content Distribution and Delivery

   • Video clips and audios are stored in Google Cloud Storage.
   • The user receives a push notification with a link to the summary.
   • Cloud CDN delivers the content quickly in the mobile app.

👥 Collaborators and Authors

• Hugo Trujillo: Project Manager
• Uriel Contreras: Lead Backend Developer
• Miguel Soto: Lead Frontend Developer
• Adrian Rocha: AI Specialist

Built With

• eas
• eventarc
• expo.io
• express.js
• firestore
• google-cloud
• google-pubsub
• google-schedule
• google-tasks
• node.js
• react-native
• redis
• redux
• spanner