Back to Search
Sophisticated Scoring v2.0

How Scoring Works

🌉 San Francisco

A rigorous analysis of building quality using 11 weighted dimensions and official SF Department of Building Inspection records.

60K+violations
11dimensions
154K+buildings
98.5max score

The Algorithm

Weighted composite with anti-clustering

Final Score =
Σ(Categoryi × Weighti)
÷ Total Weight
+ Neighborhood Adj.
+ Fingerprint
→ Compress if >92
(Safety×3.5 + Maintenance×3.0 + Track×3.0 + Heat×1.8 + Pest×1.8 + Watch×1.8 + Eviction×1.2 + Seismic×1.0 + Reno×0.8 + Transit×0.6 + Internet×0.5)
3.5×
Safety
Highest priority
3.0×
Maintenance
Building upkeep
3.0×
Track Record
Historical data
1.8×
Heat
Heating issues
1.8×
Pest
Infestations
1.8×
311 Watch
Complaints
1.2×
Eviction Risk
Displacement
1.0×
Seismic
Retrofit status

Anti-Clustering Technology

Ensuring unique, differentiated scores

Data Skepticism

Old or large buildings with zero issues are suspicious — either under-inspected or missing data.

Pre-1940 + 0 violations → -4pts
100+ units + 0 violations → -5pts

Neighborhood Adjustments

40+ SF neighborhoods have quality adjustments from -2.5 to +2.5 based on area conditions.

Pacific Heights → +2.5pts
Tenderloin → -2.5pts

Building Fingerprint

Deterministic micro-adjustment ensures every building has a unique score — no ties.

MD5(id + year + units + coords)
→ ±0.5 point adjustment

Top-End Compression

Scores above 92 are compressed to create spread at the top. Max possible is 98.5.

if score > 92:
  score = 92 + (score-92)*0.7

Result: 4,977+ unique scores across 154K buildings. No clustering at 100. Score range: 18.8 to 94.3.

Core Principles

Time Decay

6
decay tiers

Recent violations carry more weight. 90 days = 100% weight, 5+ years = 5% weight. Fixed issues fade but never fully disappear.

Open vs Closed

open weight

Open violations have 3-8x more impact than closed ones. Current problems matter most.

Per-Unit Normalization

÷
units for fairness

Large buildings naturally have more violations. We normalize by unit count for fair comparison.

Catastrophic Caps

20
minimum floor

Extreme cases get hard caps: 1000+ violations → max 20. 80+ open violations → max 20.

Max Score Caps

98.5
absolute maximum

Any violations? Max 97.5. Many violations? Lower caps: 50+ violations → max 85.

SF Climate Factor

1.8×
heat weight

SF's mild climate (50-70°F) means heat issues are weighted 1.8× vs 2.5× for NYC/Boston.

🌉

San Francisco-Specific Data

Unique factors integrated into scoring

Rent Control Status

Buildings built before June 13, 1979 are subject to SF rent control.

Seismic Retrofit

Soft-story buildings must retrofit — tracked with dedicated scoring dimension.

Eviction History

Owner Move-In (OMI), Ellis Act withdrawals — integrated into eviction risk score.

Fire Inspections

SF Fire Department inspection history contributes to safety score.

SF Rent Control: Buildings constructed before June 13, 1979 are generally subject to San Francisco rent control ordinance. We flag these buildings so renters know their potential protections. Costa-Hawkins exempts single-family homes and condos.

Grade Thresholds

A
90-98.5
B
80-89
C
70-79
D
60-69
F
0-59

Maximum possible score is 98.5 — no building can reach 100. This ensures meaningful differentiation at the top.

11 Scoring Dimensions

Tap any category to see details

SF vs NYC Scoring

Key differences in methodology

🗽

NYC (18 Dimensions)

  • • HPD violations with Class A/B/C severity system
  • • 311 complaint integration with serial-complainer detection
  • • Landlord portfolio analysis across 70K+ landlords
  • • Elevator reliability tracking for high-rises
  • • Heat violations critical (harsh winters) — 2.5× weight
  • • Rent stabilization tracking via DHCR
🌉

SF (11 Dimensions)

  • • DBI complaints with Notice of Violation data
  • • Sophisticated anti-clustering (fingerprints, compression)
  • • Seismic retrofit status (unique soft-story tracking)
  • • Eviction history (OMI, Ellis Act integrated into score)
  • • Heat violations weighted 1.8× (mild climate)
  • • Neighborhood quality adjustments (40+ areas)

Why the difference? NYC has 12M+ detailed housing violations with granular classification. SF DBI data is comprehensive but structured differently. Both systems now use sophisticated anti-clustering to ensure score differentiation — no perfect 100 scores, unique fingerprints for tie-breaking, and top-end compression.

Try It Now

See Your Building's Score

Enter any San Francisco address to get a comprehensive building report with our 11-dimension scoring.

Search Any SF Address

Related

Last updated: 2026-01-09Data sourcesHow scoring worksRenter checklist