← All professional work

Professional work

Implemented contribution

ShotLink Recording Viewer

A focused workbench integration that connects golf-shot metadata with Azure broadcast recordings.

My role

Workbench integration contributor

Focus

API integration / Desktop tooling / Broadcast analysis

Read the story
ShotLink recording browser with player, stroke, club and event metadataView full size ↗
Actual Qt viewer rendered with offline sample records. The Azure account, players and events are demonstration values.
01

Search by golf-event context

02

Correlate shots and recordings

03

Download directly into the workbench

Built with

  • Python
  • Qt
  • Azure
  • Git

Context

A recording filename is rarely how someone remembers a shot. Analysts and support staff think in terms of a tournament, round, hole, player and stroke.

My contributions

  • Recording browser: list and navigate Azure broadcast captures inside the existing workbench.
  • Metadata indexing: use ShotLink context to associate events with stored recordings.
  • Matching and caching: correlate event identifiers with timestamp fallbacks and reuse fetched metadata.
  • Desktop integration: connect selection, download progress and import to the established analysis workflow.

Design decisions

The viewer is an extension of FusionWorkbench, so it follows the host application’s interaction model. Asynchronous work and Qt signals keep network operations separate from the desktop window.

Golf identifiers also need domain-aware handling. Tournament-code differences, round context and incomplete metadata belong in the integration layer, with tests around parsing and matching behavior.

Project overview

This was a scoped contribution within the much larger VisionDevelopment repository. I added the browser and indexing helpers, integrated them with workbench window management and contributed regression fixes for matching and tournament codes.

The purpose is a shorter route from a recognizable shot to the recording someone needs to inspect. It joins existing services and analysis tools; it does not generate the source radar data or implement the surrounding vision engine.

The integration is present in the local contribution history. Its value is the workflow it enables: use familiar event metadata to locate a capture, then continue analysis in the tool the team already uses.