Built archive recovery onboarding manual for spatial research

• Day: 2026-04-29
• Time: 10:40 to 10:50
• Project: Dev
• Workspace: WP 1: Strategic / Growth & Development
• Status: Completed
• Priority: HIGH
• Assignee: Matías Nehuen Iglesias
• Tags: Archive-Recovery, Docusaurus, Documentation, Spatial-Data, Onboarding, Validation

Description

Session Goal

Reframe a scattered FCV spatial research archive into a structured onboarding and recovery system that preserves continuity, clarifies folder authority, and supports future reuse of the recovered work.

Key Activities

• Defined the archive as a multi-layer research system rather than a flat file dump.
• Separated the main 2023_Duke pipeline from the spatial_data product store and from older legacy/salvage folders.
• Proposed a Docusaurus-based documentation architecture to make the archive legible to collaborators.
• Designed multiple documentation variants, converging on a compact onboarding surface with pages such as:
  • Start here / orientation
  • Project or archive map
  • Dataset inventory
  • Notebook guide
  • Recovery / next steps
• Established an evidence-first workflow for validating what can be documented immediately versus what still needs file inspection.
• Outlined shell-based inventory and classification steps to gather compact evidence for folder mapping and downstream markdown generation.
• Clarified that matching/regression are downstream empirical components, not the whole project, and should be treated as one vertical within a broader archive.

Achievements

• Produced a coherent archive taxonomy that distinguishes:
  • the canonical 2023 pipeline,
  • reusable spatial feature products,
  • and legacy reconstruction material.
• Identified the archive map as a navigation artifact for onboarding, not just a file index.
• Established a documentation-first recovery sequence: map folders, write orientation docs, validate datasets, then decide what to reuse, rerun, or rebuild.
• Highlighted the highest-priority unresolved gaps, especially canonical geography definitions, DHS output selection, and the missing investment/project exposure layer.

Pending Tasks

• Inspect the filesystem to confirm top-level folder roles and evidence for the archive map.
• Validate dataset families and notebook metadata before expanding documentation.
• Recover or reconstruct the missing treatment/investment preprocessing layer.
• Decide which empirical outputs are reusable as-is versus which need reruns or rebuilds.
• Finalize the Docusaurus page structure and sidebar/document-ID wiring once the evidence inventory is complete.

Evidence

• source_file=2026-04-29.sessions.jsonl, line_number=1, event_count=0, session_id=3b23b2d60623e149bc93dce235a9a7ab324ebf1d26794945d24cbc795878e95b
• event_ids: []