How the TIA Score works
A single 0–100 score with a letter grade, built from six independently computed layers. Nothing about it is a black box — every layer reports its own score, its weight, and the specific factors that moved it.
What goes into the number
Weightings shift by property type. An oil and gas facility carries more weight on infrastructure, incident history, and operations, and less on ambient neighborhood crime, than a residential complex does.
Each nearby crime event is decayed by distance (squared falloff to a 1,600m radius) and by age, then weighted by severity. The aggregate risk runs through a logistic curve to a 0–100 score. Violent, property, and disorder crimes are called out separately as factors.
Lighting, sight lines, landscaping, CPTED compliance, ingress/egress control, natural surveillance, and territorial definition — each weighted, with any sub-score under 50 raised as a flagged factor with specific remediation.
Access control type sets a base ceiling (from none through multi-layer), scaled by condition. Camera coverage is multiplied by camera quality, so wide coverage of poor cameras does not score as protection.
Severity-weighted incidents with time decay. Unresolved incidents count double. Average response times over ten minutes and any critical-severity incidents surface as their own factors.
Staffing, armed status, patrol frequency, response time, quarterly training hours, certified-staff percentage, documented emergency plan, and incident reporting system.
Resident and tenant safety reports, weighted by how relevant each signal is to the specific location, aggregated into a sentiment score.
The score knows what it doesn't know
Every score ships with a completeness score, a freshness score, and a confidence score, plus an explicit list of missing and stale inputs. Low confidence applies a penalty and caps the maximum achievable score — a site with thin data cannot present as an A.
When no crime data has been logged for a location, the crime layer returns a neutral placeholder flagged as a coverage gap — never a clean bill of health. A site with genuinely logged data showing no nearby events scores materially higher than one with nothing on file.
Score to grade
This model is versioned
Every score records the version that produced it, and changes ship as new versions with an effective date and a changelog — never as a silent adjustment to a model someone is already relying on.
Version history →Any score can be challenged
An assessed party has 30 days to appeal, reviewed by someone other than the assessor of record. A correction supersedes the original rather than overwriting it.
How an appeal works →See it run against your property
An audit produces the full breakdown: per-layer scores, ranked vulnerabilities, remediation actions with estimated score impact, and where you land against comparable properties.
Request an audit