MLB-EQ

MLB-EQ (MLB Equivalent) uses Google Cloud AI and MLB data to predict a prospect’s impact. With AI-driven scouting reports and projections, we bring data-backed insights into baseball’s next stars.

Comment

Inspiration

For years, I’ve used machine learning to project NBA Draft prospects, combining analytics with scouting insights. As a lifelong baseball fan, I’ve always wanted to apply a data-driven approach to MLB scouting. Inspired by classic Baseball America and MLB Prospect reports, this hackathon was the perfect chance to build MLB-EQ, bringing AI-powered analysis to the future of scouting.

What it does

MLB-EQ (MLB Equivalent) uses Google Cloud AI and real MLB data to predict a prospect’s future impact. It focuses on key scouting criteria, including:

  • AI-Generated Scouting Reports – Uses Gemini models to assess hit, power, speed, arm, defense, and pitching tools on the 20-80 scouting scale.
  • IMAGEN Skill Icons – AI-generated visuals highlight player strengths and unique traits.
  • Interactive Prospect Search – Quickly find and evaluate MLB prospects with AI-driven insights.
  • Career Projections – Predicts best-case and likely MLB outcomes based on historical comparisons.
  • MLB Equivalent Stats – Shows past-to-future player value adjusted for league, park, and era factors.

How we built it

  • Google Cloud Run – Scalable backend hosting.
  • Vertex AI & Gemini Models – AI-generated scouting reports & projections.
  • Google Cloud Storage – Fast access to MLB datasets & AI outputs.
  • Flask & Python – API for player search, stats, and AI reports.
  • Chart.js & JavaScript – WAR projections & historical data visualization.

Challenges we ran into

  • Figuring out the right AI prompts – Took a lot of trial and error to get scouting reports that felt accurate.
  • IMAGEN Baseball Cards – Generating structured, consistent baseball card visuals with AI was harder than expected. After multiple prompt iterations, I had to put this idea on hold.
  • Choosing a Single Focus – Explored computer vision for extracting Statcast data from videos, but ultimately narrowed the project to scouting reports.
  • Familiarizing with Google Cloud – Learning the features and best practices for Google Cloud Run, Storage, and Vertex AI (chat-based Gen AI assistance was a huge help).

Accomplishments that we're proud of

  • Built MLB-EQ in just one week.
  • Created a fully AI-powered scouting platform with real MLB data.
  • Deployed on Google Cloud Run, making it scalable and efficient.

What we learned

  • Google Cloud Run makes it fast & easy to get a web app live.
  • Vertex AI is powerful for building an ML pipeline for scouting insights.
  • Experimented with computer vision for another project idea (to be revisited later).

What's next for MLB-EQ

  • MLB Equivalent Translations – Incorporate league, park, and era adjustments for better comparisons.
  • Future WAR Projections in the UI – Already built using Vertex AI offline, needs to be surfaced in-app.
  • More Ways to Explore Prospects – Add top prospect lists, search filters, and comparison tools.
  • Top Career Trajectory Matches – Show which MLB players had the most similar development paths.

MLB-EQ is just getting started—bridging AI and baseball scouting for the next generation of player evaluation. ⚾🚀

Built With

Share this project:

Updates