Why we are building this, and what we are trying to get right.

What we are trying to do

Most neighborhood safety information in Ethiopian cities lives in two places: formal police reports (incomplete, unevenly distributed, not publicly accessible) and informal community knowledge (rich, real-time, but private and unstructured).

Gorebet exists in the gap. We give communities a structured way to record what they observe, give institutions a governed way to receive that signal, and give researchers an honest view of what the data can and cannot show.

We are explicit that we are not a crime-fighting tool. We do not promise to reduce crime rates. We provide the information infrastructure for communities, institutions, and services to make better-grounded decisions.

How we approach it

Community first, institution second

The Gorebet app is built for residents first. They decide whether to report, whether to use their name, and what to share. Institutional partners receive aggregated data from that community activity — not direct access to individual reports without governance agreements.

Honest about what we measure

Gorebet data reflects what community members chose to report. It is not a census of incidents that occurred. Low reporting rates may mean low crime, low platform adoption, or low community trust — we cannot distinguish these from the data alone. We label coverage gaps explicitly.

Fairness as a design requirement

We commit to annual fairness audits comparing how our platform performs across language groups, approximate demographic groups, and woreda socioeconomic levels. We have not validated all of our measures. We publish what we have and have not done on the methods page.

Privacy by design

Anonymous reporting is end-to-end: no identity is stored with anonymous reports. All data is stored in Frankfurt, Germany. No advertising technology. No cookies on this website. Details on the trust & privacy page.

No enforcement orientation

We explicitly do not recommend that Gorebet data drive enforcement decisions without human review, institutional governance, and bias auditing. We tell this to every government and operator partner. Incident data is a supplement to — not a replacement for — field investigation.

The research tradition we build on

Gorebet's design is grounded in criminological and sociological research: collective efficacy theory (Sampson, Raudenbush & Earls 1997), routine activity theory (Cohen & Felson 1979), time-geography (Hägerstrand 1970), environmental criminology (Brantingham & Brantingham 1981; Sherman et al. 1989), and computational social science methodology (Salganik 2017; Barocas, Hardt & Narayanan 2019).

We also take seriously the critiques of this tradition: Harcourt (2001) on the empirical weaknesses of broken-windows policing; Sampson (2012) and Lum & Isaac (2012) on how neighborhood information systems can reinforce spatial inequality; and the community-of-practice consensus that bias in training data produces biased outputs at scale.

Who builds Gorebet

Gorebet is built by Gorebet Inc., a software development company based in Addis Ababa, Ethiopia. We are a small team building civic-technology infrastructure for Ethiopian cities.

We are not affiliated with any law enforcement agency, government body, or political organization. We work with government agencies as partners — they receive governed access to aggregated data. They do not control the platform, set reporting policies, or have access to individual reporter identities.

Get in touch

For research collaborations, partnership inquiries, press requests, or general questions, use our contact form or write to us directly.