Methodology and editorial standards
These were written before SafeNaija published anything, and they are published here so anyone can hold us to them. A standard written in advance is a commitment. One written afterwards is an excuse.
Sources
Where the data comes from
| Source | What it provides | Licence |
|---|---|---|
| UCDP Georeferenced Event Dataset Uppsala Conflict Data Program |
Historical incidents, located to village level. Covers organised violence above a fatality threshold — it misses ordinary crime and non-fatal abduction, so it is a floor, not a full picture. | CC BY 4.0 — republished with attribution |
| Nigerian news organisations Premium Times · Vanguard · Punch · Daily Trust · HumAngle · The Cable |
Current incidents. We record the facts of an event — date, place, type — and link to the original. We never reproduce an article. | Facts are not copyrightable under s.3 of the Copyright Act 2022. Every record credits and links the outlet. |
| Community reporters | Incidents on roads before or instead of press coverage. Never published on one report alone. | Submitted anonymously; no identity is stored |
If your organisation's work appears here and you would like the attribution changed, or removed, write to corrections@safenaija.ng.
Editorial standards
Twelve rules, no exceptions for speed
We publish only what has already been published
Every incident cites a named Nigerian news organisation with a link to the original, or has been corroborated by two independent reporters. SafeNaija is an index of what was reported, not a source of new allegations.
We never publish an ongoing incident
No live hostage situations, no active attacks, no ongoing security operations. Every record is held before publication.
We publish no operational detail
No movements or positions of security forces, convoys, checkpoints or deployments. No response times.
We never name who was responsible
Incidents are described by what happened, never by who is alleged to have done it. No group, ethnicity, religion, community or nationality is ever named.
We never name people, businesses or premises
No individual, business, compound, hotel, school or residence is identified. Locations are given as a road or an approximate area.
We publish no casualty figures of our own
Where a source reports a number, we cite it as theirs. We keep no death toll, counter or index. A traveller choosing a route does not need a national body count.
We do not rank states, governments or agencies
SafeNaija shows roads and areas. It produces no league tables of states, local governments, governors or security agencies.
We say nothing about government performance
SafeNaija records what happened, not what any authority did or failed to do. Accountability journalism is essential work; it is not our work, and mixing the two would compromise both.
We are a road and travel safety service
Not a conflict monitor, not an intelligence service, not a watchdog. Our purpose is to help a person choose a safer road.
We correct quickly and publicly
Disputes are answered within 24 hours. Anything we cannot substantiate comes down while we review it, and every correction is logged with its reason.
Reports can describe danger only
There is no mechanism to mark a road, area or route as safe or clear, and there never will be.
We never promise safety
SafeNaija never describes a route as safe. It reports how many incidents have been recorded and when. Nothing we publish should be read as a guarantee of any kind.
The hardest question
"Doesn't this help the people causing the violence?"
It is a fair question and we would rather answer it here than be asked it for the first time in public.
A map of where incidents have occurred is, in principle, as useful to someone planning an attack as to someone avoiding one. That tension is real and cannot be fully removed. Four things reduce it:
- Nothing here is new information. Every incident on the map was already published by a Nigerian newspaper, usually with more detail than we carry. We add no fact that was not already in the public record.
- Locations are coarse. A grid cell roughly a kilometre across is useful for choosing a road and useless for choosing a target.
- Nothing is live. Time passes before anything appears. There is no real-time feed of where security forces are or are not.
- Safety cannot be reported. The single most dangerous feature such a map could have — the ability to mark a road clear — does not exist in our system and cannot be added without rebuilding it.
We think a traveller deciding between two roads is better served by a dated, sourced, deliberately imprecise record than by a rumour in a WhatsApp group. That is the judgement the project rests on, and we will revisit it if the evidence changes.
Limits
What this data cannot tell you
- Absence of markers is not safety. It means nothing has been reported. Roads nobody reports on look identical to roads where nothing happens.
- Coverage is uneven. Incidents near cities and on well-travelled roads are reported more often than those in places with no press presence and no mobile signal.
- We cannot predict. SafeNaija shows what was recorded. It does not forecast, and no part of it should be read as a prediction.
- Historical data is partial. UCDP codes organised violence above a fatality threshold, so much ordinary crime and non-fatal abduction is simply not in it.