Transparency & Methodology
We believe in full transparency about how SafeNSound works. This page explains our data sources, scoring methodology, and limitations.
Our Data Sources
FBI NIBRS Crime Data
Sex offenses and kidnapping/abduction crimes aggregated by state
Limitation: Aggregated by state only; not all agencies report to NIBRS
Sex Offender Registry (Aggregated)
Estimated registered sex offenders per 100,000 residents by state
Limitation: Estimates only; actual numbers may differ; aggregated at state level
Missing Persons Reports
Missing persons and kidnapping cases by county
Limitation: THIS IS SAMPLE DATA for demonstration. Not actual missing persons statistics.
USDA/Census SAIPE 2023
County-level poverty rates from the Small Area Income and Poverty Estimates program
Limitation: Poverty is a correlative indicator, not a direct safety measure
IPEDS College Data (NCES)
Accredited colleges and universities — campus presence improves area safety scores
Limitation: Safety benefit is modeled, not measured. Based on campus security presence and infrastructure.
NCES School Data
Middle and high schools with location coordinates
Limitation: Does not include elementary schools or private institutions not in NCES
ICE Detention Centers
Locations of ICE detention facilities and documented protest sites
Limitation: Sample data; not a comprehensive list of all facilities
Commercial Locations
Big-box retail, fast food chains, and Starbucks locations
Limitation: Small sample; not representative of all commercial areas
Parks & Playgrounds
National parks and playground locations
Limitation: Small sample; focused on national parks
Florida Sex Offender Registry
Individual registrant records aggregated to county level
Limitation: Only covers Florida; no individual data is displayed in the application
How We Calculate Safety Scores
Each location receives a composite safety score from 0 (safest) to 100 (most dangerous). The score is a weighted sum of seven individual factors, each normalized to a 0-100 scale.
Factor Weights
Normalization Method
Each factor is normalized using min-max scaling:
The composite score is then:
What the Scores Mean
Sex Offender Registry Data Gaps
NSOPW (the National Sex Offender Public Website) is not a database — it is a federated search portal that queries each state's public-facing registry in real time. At least 12 states use risk-based tiered systems that exclude lower-tier offenders from their public websites. SafeNSound applies statistical correction factors to account for these gaps, but users should be aware that the underlying data in these states is inherently incomplete.
| State | Est. Total | Publicly Listed | % Hidden | Public Disclosure Rule |
|---|---|---|---|---|
| OR | 33,400 | 1,266 | 96.2% | Level 3 only |
| MN | 19,821 | 1,287 | 93.5% | Level 3 only |
| VT | 1,275 | 400 | 68.6% | Court-determined; high-risk/SVP/assault |
| WA | 20,000 | 6,500 | 67.5% | Level 2 + 3 |
| NJ | 14,742 | 5,200 | 64.7% | Tier 3 + qualifying Tier 2 |
| AR | 18,000 | 7,200 | 60.0% | Level 3 + Level 4 (SVP) only |
| RI | 3,400 | 1,370 | 59.7% | Level 2 + 3 |
| MA | 11,506 | 5,665 | 50.8% | Level 3 + post-7/2013 Level 2 |
| NY | 42,611 | 27,022 | 36.6% | Level 2 + 3 |
| AZ | 17,000 | 11,000 | 35.3% | Level 2 + 3 |
| CA | 105,000 | 75,000 | 28.6% | Tier 2 + 3; Tier 1 excluded (SB 384) |
| NV | 7,300 | 5,500 | 24.7% | Tier 2 + 3; Tier 1 w/ child victim only |
Sources: DOJ SMART Office, state BCA/DOC/DPS reports, NJ Courts Megan's Law reports, MA State Auditor, NY DCJS, SafeHome.org. An estimated 100,000+ registered sex offenders are invisible to NSOPW across these states. Only 18 of 50 states have substantially implemented SORNA (Sept 2024 SMART Office Progress Check).
Important Disclaimers
•SafeNSound scores are statistical estimates for educational purposes only, not guarantees of safety or predictions of danger.
•A low score does not mean an area is unsafe; a high score does not mean imminent danger exists.
•This tool does not identify, target, or track any individuals. All sex offender data is aggregated at the county/state level.
•No individual names, photographs, or addresses are displayed in this application.
•Some data sources contain sample/mock data (clearly labeled) and may not reflect current conditions.
•The time-of-day modifier in route analysis is a general heuristic based on criminological research, not actual time-specific crime data for any specific location.
•This tool is not a substitute for personal judgment, local knowledge, or law enforcement advice.
•SafeNSound does not collect, store, or share any user location data.
•Environmental exposure layers (nightlife, social venues, commercial areas) are informational only and are not factored into danger scores.
Known Limitations
Geographic coverage gaps: Point-level sex offender data covers all 50 states via NSOPW, but at least 12 states hide lower-tier offenders from public registries (see Registry Data Gaps above). We apply statistical correction factors to rankings and hex scores, but cannot place missing offenders at specific locations.
Sample data: Missing persons data is explicitly mock/sample data created for demonstration purposes. Commercial location data covers approximately 270 locations, not the full national inventory.
No real-time data: All data is static as of the dates indicated. Crime patterns change over time and may not reflect current conditions.
Correlation vs. causation: Poverty rate is included as a correlative risk factor based on research. It does not imply that poverty causes crime or that impoverished areas are inherently dangerous.
Scoring model limitations: The weighting system reflects general criminological research priorities but is not validated by an expert panel. Different weightings would produce different scores.
Bias & Fairness
We acknowledge that any data-driven scoring system can reflect and potentially amplify existing biases in the underlying data.
Crime statistics may be influenced by reporting practices, enforcement patterns, and systemic factors that disproportionately affect certain communities. A higher score does not necessarily mean an area is more dangerous -- it may reflect differences in data collection and reporting.
SafeNSound is designed to inform, not to label. We do not target any individuals, groups, or protected classes. All insights are derived from environmental factors and publicly available aggregate datasets.