Route safety
How anonymous matatu route signals can improve safety without exposing riders
Route accountability should use minimized, aggregated, consented signals, not private evidence or exact journeys.
Review status: Route-data governance and safeguarding review required before production publication
A route signal should never contain a survivor’s private evidence
A rider may want an incident to contribute to a broader picture of safety without sharing a statement, audio recording, photograph, name, or exact journey. SafeRide separates the private incident record from the smaller set of fields considered for route-level analysis.
That separation should be enforced in the data model and visible in the interface. A route signal must be a separate, voluntary choice. It should never be created automatically because someone drafted a report or asked for support.
Collect the minimum detail needed to understand a pattern
Useful route information may include a broad corridor, a time window, and a high-level incident category. It should exclude names, contact details, raw media, exact pickup or destination points, and any combination of details that could identify a rider.
Data minimization also means declining information that is interesting but unnecessary. Each field should have a documented purpose, retention period, access rule, and deletion path before the pilot treats it as part of a public-safety dataset.
Do not publish route scores until the sample is large enough to be responsible
A small number of reports can create a misleading impression about a route, operator, neighbourhood, or time of day. A responsible index needs minimum sample thresholds, clear time windows, uncertainty notes, duplicate controls, and a process for reviewing bias or coordinated misuse.
The absence of reports cannot prove that a route is safe. Differences in phone access, awareness, willingness to report, and partner coverage can all affect the data. Public explanations should make those limits as visible as any score or trend.
Route information matters only when partners act on it
A dashboard is not an outcome. Transport operators, survivor-support organizations, civic partners, and regulators need a documented process for reviewing patterns, choosing a proportionate response, and reporting what changed without exposing individual riders.
SafeRide’s pilot should measure whether partners understand the signal, whether an action follows, and whether riders experience the change as safer. Until that evidence exists, route pages should describe the method and limitations rather than claim verified safety rankings.
Choose your next step.
Read the Route Safety Index method
See the proposed safeguards, thresholds, partner roles, and limits of route-level reporting.
ContinueCompare private and anonymous data
Understand why personal evidence and community route signals need separate consent decisions.
ContinueWork on a responsible pilot
Review the roles available for operators, survivor-support organizations, researchers, and funders.
ContinueSources and further reading
- SafeRide in the UNICEF Femtech Ventures cohort— UNICEF Office of Innovation