Second Meeting Record
Meeting Description
Date / Time: 29 March 2026 (evening)
Location: Online discussion after instructor feedback
Attendees: Xi Zixun, Wang Zhuoran, Jin Qinhao
Purpose
This meeting recorded the team decision to pivot from the earlier mixed recovery direction into one shared tourism time-series workflow for the final project.
Key Decisions
- Use
data/raw/visitor_arrivals_full_dataset.xlsxas the shared arrivals backbone. - Keep
data/raw/tourism_update.xlsxonly as supporting tourism context for EDA, CDA, and forecasting interpretation. - Reframe the final app around three modules:
- Time Series Visual Analysis
- Time Series Clustering
- Forecasting
- Treat country-level monthly arrivals as the common analytical target across modules.
- Remove the older standalone modelling materials so the final project only exposes the three shared time-series modules.
Work Split After Pivot
- Xi Zixun: visual analysis rewrite, proposal intro/data/research questions, shared arrivals framing.
- Wang Zhuoran: clustering prototype, clustering Shiny module, storyboard, app shell integration.
- Jin Qinhao: forecasting prototype, forecasting Shiny module, user guide, package audit, smoke test updates.
- Shared integration work: navigation cleanup, deployment path, meeting records, and final QA.
Action Items
- Build processed country-arrivals backbone files from the new full workbook.
- Align website navigation with the final three-module structure.
- Update proposal, storyboard, and user guide so they describe the same app contract.
- Deploy the static website separately from the Shiny app.
Expected Deliverables
- One Quarto website with proposal, prototypes, user guide, and meeting records.
- One Shiny app with visual analysis, clustering, and forecasting modules.
- Shared scripts that can regenerate the processed arrivals backbone when the raw workbook changes.