Island Watch
Case studyThe map is only the visible layer. The real work is source ingestion, regional normalization, freshness labels, and calm public-status UX.
Public-source ingestionGeospatial decision surfaceFreshness and source attribution
I build calm tools that turn messy real-world inputs into clear decisions.
Product engineer in the Comox Valley building browser-native analysis tools, civic status surfaces, local guides, and manager operations systems. The work is technical; the point is practical: help people understand what matters and what to do next.
Product proof
InputSignalMove
Public feedsIsland WatchCheck regional status
Engine outputChessIQReview and train
Event sourcesIsland HappeningsChoose what is worth doing
Manager ritualsManager BinderPreserve the handoff
Working habit
Collect the messy inputs, name the signal, then make the next move repeatable.
Browser-native analysis
Stockfish WASM, workers, progressive feedback
Operational data systems
Ingestion, normalization, source freshness
Local-first persistence
IndexedDB, routeable state, repeat-use workflows
Trust-oriented UX
Source links, clear states, decision surfaces
Four domains, one product habit
Public feeds, chess analysis, event listings, and manager handoffs all start noisy. The work is to preserve trust, expose what matters, and make the next step clear.
01Public utility
Island Watch
Regional awareness for public updates people need to verify quickly.
- Signal
- What changed, where it applies, how fresh it is, and where the official source lives.
- Hard part
- Official updates are useful but scattered across sources, formats, jurisdictions, and freshness states.
Proof
- Scheduled public-source ingestion
- Normalized alert model
- Source freshness and trust states
02Browser-native analysis
ChessIQ
Browser-native chess review that turns engine output into trainable missed patterns.
- Signal
- Missed candidate moves, recurring calculation habits, and positions worth practicing again.
- Hard part
- Stockfish output is powerful, but raw engine lines can bury the learning signal inside evaluations and move lists.
Proof
- Stockfish WASM + web workers
- IndexedDB study persistence
- Progressive analysis with runtime safeguards
03Curated local guide
Island Happenings
A curated local guide for what's actually worth doing nearby.
- Signal
- Dates, locations, cost, planning notes, and the few choices that deserve attention this week.
- Hard part
- Event discovery often optimizes for volume when people need timing, place, source confidence, and practical context.
Proof
- Source-linked listings
- Brief-first filters
- Submission and correction paths
04Operations systems
Manager Binder
Operational memory for managers keeping recurring work on track.
- Signal
- The next responsibility, the decision history, and the pattern behind daily operations.
- Hard part
- Manager work leaks into memory, message threads, spreadsheets, and rituals that only make sense to the person who last did them.
Proof
- Local-first scheduling module
- Templates, role fit, and review flows
- Continuity over tribal knowledge
Where the hard parts show up
The strongest projects are more than polished interfaces. They carry ingestion, persistence, normalization, source trust, and decision-surface work that has to hold up in repeated use.
ChessIQ
Case studyA browser-native study workflow: engine analysis, persistent review, and training generated from real missed positions.
Stockfish in-browserIndexedDB-backed study stateOwn-game training loop
Latest notes from the hard parts
These pieces unpack the decisions behind the work: what to hide, what to expose, when to trust a signal, and how technical systems become usable products.
All notesBuilding a Chess Coach That Shows Its Work
ChessIQ's Patterns page was rebuilt around a simple trust rule: if the app says something keeps happening, it has to show the games, moves, confidence, and training path behind the claim.
How ChessIQ Turns Statistics Into a Player Portrait
ChessIQ's statistics surface turns local reviewed-game history into a living five-factor player portrait while keeping confidence, sample size, and visual restraint part of the contract.
The hard part of Island Watch wasn't the map
Island Watch looks like a map dashboard, but the harder work is source freshness, alert normalization, route relevance, official handoff, and spatial trust.