tiger-history-deep

name	tiger-history-deep
description	Deep revision history analysis for specific OSM ways — fetches full version history from OSM API v0.6, identifies import bot vs human edits, classifies review status with confidence scores.
when_to_use	User mentions a specific way ID, asks about edit history, checks if a way was reviewed, discusses INCONCLUSIVE status, or wants to investigate a specific street
allowed-tools	Read Grep Glob Bash(python *) Bash(curl *)
argument-hint	[way-id or street-name]
arguments	["target"]

Deep Revision History Analysis

Investigate: $target

Why this matters

The core innovation of this project (per Minh Nguyen's feedback) is that tiger:reviewed=no is unreliable. Most mappers don't remove the tag after correcting data. Deep history analysis is the only way to determine if a way has been meaningfully reviewed since the 2007 TIGER import.

Analysis tiers (from history_filter.py)

Tier 1: Fast check (from out meta data already in scan results)

• version == 1 and user == "DaveHansenTiger" → definitely UNREVIEWED
• Last edited recently by a human mapper → likely REVIEWED
• Ambiguous → proceed to Tier 2

Tier 2: Full history (OSM API call)

• GET /api/0.6/way/{id}/history — all versions of the way
• GET /api/0.6/node/{id}/history — for nodes along suspicious ways
• Analyze what changed between versions:
  • Was the change to geometry (node positions moved)?
  • Was the change to tags (highway, oneway, name corrected)?
  • Was it incidental (only tiger:* tags added/removed)?
  • Was oneway explicitly set by a human or inherited from import?

Output per way

• review_status: UNREVIEWED | LIKELY_REVIEWED | INCONCLUSIVE
• review_confidence: 0.0 – 1.0
• Edit timeline: who changed what, when

Known import bot accounts

These edits do NOT count as human review:

• DaveHansenTiger — original 2007-2008 TIGER import
• Abbreviation-expansion bot (December 2012) — automated name cleanup
• January 2010 node-tag cleanup — removed source=tiger_import_dch_v0.6_* tags from nodes

Rate limiting

• OSM API allows ~1 request/sec for history endpoints
• Cache responses to ~/.config/osm/history_cache/ with 7-day TTL
• For bulk analysis, the scan pipeline's --skip-history flag exists for speed

Current scan results

!for d in osm-audit-*/; do f="$d/scan-results.json"; [ -f "$f" ] && echo " ${d%/}: $(python -c "import json; print(json.load(open('$f'))['summary_stats']['total'])") ways"; done 2>/dev/null || echo " (no scans yet)"