← Projects

Sports tracking & analysis

SportsCarpet

A sports workspace for provider data, league preferences, saved slips, recorded activity and analysis, with practice credits kept separate.

My roleProduct design, application development and provider integration.
Reconstructed SportsCarpet Android pilot screens for practice wallets, fixtures and connected analysis.
Pilot presentation · sample dataFull size ↗

Overview

Project overview

SportsCarpet is a locally tested pilot for following sports information, connecting providers, organising saved slips and recorded activity, scanning inputs and reviewing analysis. Practice SC balances and recorded bets stay separate. The pilot also includes league preferences, cached provider data and controls over updates.

  • Provider-connected sports data
  • League preferences
  • Wallets, slips & recorded activity
  • Scan & analysis

Engineering contribution

My contribution

  • Connected sports-provider APIs with league preferences and analysis workflows.
  • Built a workspace spanning My Wallets, My Slips, My Bets, Scan and My Analysis.
  • Separated practice credits from recorded activity and tested the pilot locally.

Technology & scope

  • Android
  • Provider APIs
  • Assistant workflows
  • Local state / caching

Technical details

Technical implementation

Connected sports data, reviewable analysis and separate records, with explicit readiness checks at each step.

01

Provider configuration and caching

Provider defaults and league preferences supply the initial workspace. Update controls allow requests to be paused, and cached information remains available when a provider allowance is exhausted.

02

Record and balance separation

My Wallets, My Slips, My Bets, Scan and My Analysis organise the application. Practice SC credits remain separate from recorded activity and its history.

03

Assistant proposals

Assistant connections check the available account, model and capabilities. Structured proposals keep evidence visible and do not write directly into the application ledger.

Workflow

Assistant analysis workflow

  1. 01

    Connect

    Choose a provider, account and model.

  2. 02

    Context

    Review the fixture and source information.

  3. 03

    Propose

    Receive a structured assistant response.

  4. 04

    Review

    Decide what belongs in the workspace.

Proposals do not write directly into the ledger. Practice credits stay separate from recorded activity.

Account connection details

Fixture data and analysis

In the pilot, the connection workflow prepares the account, model and source context before an analysis request. The user reviews the result before keeping it.

SportsCarpet pilot connection screens showing ChatGPT sign-in steps and a sample fixture carried into reviewed analysis.
Pilot connection workflow · sample fixture · Full size ↗

Verification

Build & verification

AI-assisted implementation is checked through state-transition tests and screen walkthroughs before a pilot release.

Verification examples
Account and model readiness checks
Account recognition, model selection and the connection check have separate states. This avoids showing a misleading reconnect prompt when only model discovery needs attention.
Practice-credit and ledger checks
Pilot checks cover imported records, ledger transitions and the separation of practice credits from recorded activity. Live provider and device checks remain release gates.
Next projectMy Portfolio