See the portfolio
Load the submitted locations and measure each one using the same analysis settings.
Physical-security decision intelligence · Toronto
Know which locations need security attention first.
Locivra tells you where to inspect first—and shows why the data says so.
- Historical reported crime
- Decision support
- Not crime prediction
Locivra applies one documented method across 5–20 locations, ranks where deeper security review should start, and shows the evidence and uncertainty behind every position.
Built by Alex Babb, a final-year Criminal Justice and Public Policy student at the University of Guelph. View Alex’s profile.
Prefer not to email? Message Alex on LinkedIn.
Illustrative portfolio · 10 locations
Security review queue
01
Review first
Queen–Spadina site
Multiple elevated contributors · verify late-hour theft signal
02
Validate
Yonge–Eglinton site
Mixed recent direction · internal incident check required
03
Monitor
North York West site
Lower relative exposure · no immediate escalation signal
One methodThe same documented logic is applied across every submitted site.
One review orderLimited security attention is focused where the evidence is strongest.
One decision recordDrivers, limits, validation questions and next actions stay attached.
Practitioner-shaped: the method has been refined through feedback from people working in CPTED, public safety and retail loss prevention.
The decision, not the dataset
Public data is the input. The product is the review process.
General AI can summarize a public dataset. Locivra is built to apply consistent assumptions across an entire portfolio, expose what changed the result, and turn the output into a reviewable next step.
A one-off prompt
Provides an answer.
Useful for exploration, but difficult to repeat or defend as a portfolio decision.
- Assumptions may change between requests
- Sites may not be compared on one baseline
- Uncertainty can disappear inside the summary
- No built-in validation or decision trail
The Locivra workflow
Creates a review order your team can challenge.
The result is not an automated verdict. It is a documented starting point for security, loss-prevention, risk and property teams.
- Same source, radius and scoring profile across every site
- Category drivers, trends and sensitivity shown separately
- Small samples, missing data and conflicting signals flagged
- Each priority ends with a validation question and next action
See what moved the rank
Separate raw incident volume, weighted exposure, historical direction and uncertainty.
Make the next decision
Compare the external signal with internal incidents, site knowledge and operating context.
Decision-ready evidence
A score without a “why” is noise.
Every priority needs enough evidence for a manager to understand it, question it and decide what happens next.
- Relative portfolio rank and category contributors
- Raw-count and weighted-exposure changes labeled separately
- Coverage, missing-data and low-volume warnings
- Radius and weighting sensitivity where the result may move
- A plain-language validation question for the team
LOCIVRA / LOCATION EVIDENCE RECORD
Sample record
Review order 01 of 10 · Queen–Spadina site
Why this location is first
82Relative review score
not probability
not probability
Decision summary
Several categories are elevated at once. The theft signal is the largest contributor, but mixed historical direction means internal incident and site-level validation should happen before controls are selected.
Primary contributorTheft over $5,000 exposure
Historical direction6-month rise / 12-month decline · mixed
Boundary sensitivityMaterial across tested radii
Evidence qualityCoverage recorded · low counts flagged
Next validation question: Do internal incident, loss and site-condition records support the external theft signal?
Historical reported-crime analysis. Decision support—not crime prediction, a safety certification or a replacement for site review.
A bounded validation pilot
The pilot succeeds only if it improves a real decision.
Before analysis begins, the team agrees on the portfolio, the decision being tested and how usefulness will be measured.
01
Baseline
Document the current review process, time required and existing priorities.
02
Analyze
Rank 5–20 Toronto locations and produce portfolio and location evidence records.
03
Validate
Compare the findings with internal incidents, practitioner knowledge and operating context.
04
Measure
Record where the evidence supported, challenged or changed the existing review plan.
Locivra does not claim savings, loss reduction or crime prevention without client evidence. A pilot is designed to test whether the workflow improves prioritization, clarity and resource allocation in the client’s actual process.
Founder and method
Built in Toronto for decisions Toronto teams actually face.
Locivra began with a practical question: when one team is responsible for many locations, how can it decide where limited security attention should go first?
Alex Babb · FounderFinal-year Criminal Justice and Public Policy student at the University of Guelph. Building Locivra around transparent Toronto Police Service Open Data and practitioner feedback. Connect on LinkedIn.
Locivra is validation-stage. Its method, assumptions, scope and limitations are made visible so clients can challenge the output before acting.
Data provenancePublic source, coverage period and analysis settings are documented in the output.
Reproducible methodThe same configured inputs and method are intended to produce the same result.
Visible uncertaintySmall samples, missing data, conflicting trends and sensitive boundaries are flagged.
Human validationInternal records, site knowledge and professional review remain part of the decision.
Buyer questions
What security and risk teams usually ask first.
Clear boundaries make the product easier to evaluate. These answers are intentionally direct.
What data does Locivra use?
The current Toronto analysis uses Toronto Police Service Open Data for Assault, Robbery, Break and Enter, Auto Theft and Theft Over $5,000. Every report states the data source, coverage dates, analysis radius and configured scoring profile.
How current is the information?
Locivra is historical decision support, not live incident monitoring. Each report shows its exact coverage end date so a reviewer can see how current the underlying file is before relying on the result.
What happens to client data?
A location list and a defined decision are enough to begin. Internal incident or loss records are optional validation inputs. Data-handling, access and retention requirements are agreed before any internal records are exchanged; Locivra does not currently claim a formal enterprise security certification.
Is the result auditable or a safety certification?
The output records sources, dates, settings, missing-data warnings, sensitivity checks and limitations so the analysis can be reviewed. It is not an insurance assessment, crime prediction or certification that a site is safe or unsafe.
Start with one portfolio decision