Hello World! The Fire Ants’ latest application: MLB Pitcher’s Friend
Who are the Fire Ants you ask? We are an Application Development Team formed from the Commercial Pre-Sales organization of Dell EMC, a subsidiary of Dell Technologies. We develop and deploy Cloud Native Applications, and use modern analytics frameworks to derive insight from novel data sources. Our goal? To practice what we preach: use modern infrastructure, application development platforms, and DevOps practices to provide real-world relevance to our customers. How can you, too, realize your digital future?
An Example? Pitcher’s Friend, recently released on your favorite app store, started out as an idea in a member’s mind. Harking back to his passion for baseball, and having been a college-level pitcher himself, Jason Battles, the Fire Ants fearless leader, had a vision: “Helping hitters become batters...and then sitting them back down.”
Diving deeper, here’s an excerpt from our Readme in our github repository (https://github.com/fire-ants):
· The MLB Pitchers Friend uses data analytics techniques to visualize hitting characteristics for specific Major League Baseball (MLB) players. The app uses a cloud-based, microservice architecture. The major components are described below.
Analytics Ant - analyzes pitches experienced by hitters and the outcomes. Applies the Fire Ant proprietary Hitter Val (HVAL) qualitative and quantitative scoring algorithm on a per pitch basis to determine expected outcome per pitch. Creates a comparison visualization between traditional Heat Map and HVAL and stores it in the Virtustream Object Storage for the Mobile Application. Written in R.
Queen Ant - orchestrates the aggregation (Crawler Ant) and analysis (Analytics Ant) microservices.
Crawler Ant - connects to MLB Game Day data mart and extracts observed variables on every pitch thrown and At Bat events for the targeted analysis window. Writes information to the database (Database Ant) via the API (API Ant). Written in Java.
API Ant - provides a front-end to the Database Ant and facilitates bi-directly communication (read / write). Written in Java.
Database Ant - stores all raw data provided by the Crawler Ant. Currently hosted on MongoDB.
Mobile App - presents visualizations and per-pitch analysis to the user. Written in React Native to allow for Apple and Android compatibility.
Here’s a glimpse of our current application framework:
Today, what we’ll call version 1.0 of this application is complete: the primary output within the application are per-hitter heatmaps of batting success, categorized by pitch type. These heatmaps represent the HVAL score, providing insight into optimized locations in the strike zone for pitcher success. HVAL is unique because it blends qualitative aspects of a given at-bat’s outcome, as gleaned from the MLB GameDay announcer record, and refined with simple natural language processing techniques.
Put in simple terms, HVAL aims to quantify the quality of a batting outcome- was the hit to left field sharply hit, or was it a pop fly? towards 1st base a line drive, or softly hit? These additional qualitative aspects add detail and insight beyond whether the result of a batting sequence was a hit or not, and provide greater value to a pitching coach or pitcher about where to place their next pitch.
So, what should you do next? Go download the app! Provide feedback here or through the app store, and tell us what you’d like to see next! Furthermore, we have an exciting update coming in the near future that will utilize a logistic regression model (already implemented, with front-end development near complete!) to provide tailored recommendations based on the HVAL data, so stay tuned for the next release.
Thanks for reading, and happy developing! Until the next time,