AI Hockey Analysis and Betting
Project Overview
We follow an ETL-pipeline development process with iterative delivery and tight coordination across frontend, backend, and data automation teams. The platform is built using Python and Django, with dynamic data views rendered via HTML, CSS, and JavaScript. We leverage BS4 and Selenium for web data extraction, enabling advanced hockey analytics and real-time updates. Background tasks are orchestrated using Celery and RabbitMQ to handle heavy data processing and scheduled jobs. All infrastructure is containerized with Docker and deployed to cloud environments via Terraform for consistent, scalable delivery. The site is optimized for performance and responsiveness, delivering fast, interactive insights for fans and analysts alike. Automated testing, code reviews, and CI/CD pipelines ensure stability and frequent feature rollouts. Documentation and task tracking are maintained in Notion and Linear to support efficient, transparent workflows across the team.
Technology Stack
JavaScript
Python
CSS
HTML
Selenium
Terraform
React
Application Showcase
Challenge
- Aggregating NHL data from disparate web sources (stats, schedules, news) required a reliable ETL pipeline; manual collection wouldn’t meet freshness or scale needs for AI picks.
- The product needed near-real-time updates to power betting/DFS recommendations and dashboards, plus resilient scheduling for heavy jobs.
- Frontend had to present dynamic analytics with responsive performance across devices while keeping iteration velocity high.
Our Solution
- Backend: Python/Django platform with an ETL pipeline. Web data extraction via BeautifulSoup and Selenium; background processing and scheduled jobs with Celery + RabbitMQ.
- Frontend: dynamic data views (HTML/CSS/JS) with React for interactive UI components and real-time insights delivery.
- Infra/DevEx: containerized services (Docker), cloud deployment with Terraform, CI/CD and automated testing for frequent, stable releases.
Results
- Continuous, automated ingestion replaced manual collection; ETL keeps analytics current for AI hockey picks and dashboards.
- Low-latency updates and reliable job orchestration improved data freshness and user experience for fans and analysts.
- Production-ready delivery: responsive UI, performance optimizations, and a CI/CD pipeline enabling rapid feature rollout.