RaceEdge is a public racing data-analytics platform that auto-collects race cards, odds and results from keirin and horse-racing tracks nationwide, processes them through its analytics & prediction engine, and delivers real-time predictions (◎○▲△) and odds movements.
A new-platform launch story: tackling the challenges of “scattered data, manual analysis, untrackable odds movements and prediction history that never persisted” by building in-house from scratch in 12 months.
The public racing data service provider aggregates race-card, odds and result data for keirin and horse racing, and runs a service delivering it to members as predictions and analysis. Data freshness and real-time performance are at the core of its value.
Previously, data was scattered across multiple official sites, and both collection and analysis were manual. So we set out to build, from scratch, a foundation that automates everything from data collection to prediction delivery.
Race cards, odds and results wereSplit across multiple sites, collection took significant manual effort.
Shaping and analyzing the data needed for predictionsDone manually, real-time immediacy was out of reach.
Odds movements just before the closeCouldn't be tracked in real time, delivery was delayed.
Prediction rationale and hit history wereNot accumulated, accuracy could not be verified or improved.
We built automated data collection, the prediction engine, a real-time delivery PWA and an operations CMS in-house from scratch — designing the data pipeline end-to-end, from collection to delivery.
Next.js (PWA)
React Query
Ant Design
NestJS
Puppeteer
MySQL
Redis / Bull
AWSThe crawler collects data 24/7, freeing up people.
Auto-generate predictions the moment data is collected, and deliver them instantly.
Track odds movements right up to the close, instantly, via the PWA.
Accumulate hit history to verify and improve prediction accuracy.