Methodology
This lab is an analytical tool. It does not predict the future, and it never publishes official results. Official results are declared only by Nigeria's electoral commission.
How a projection is computed
- Expected ballots in a state = registered voters × expected turnout (Scenario assumption).
- Expected valid votes = expected ballots × valid-ballot percentage (scenario assumption).
- Candidate votes = expected valid votes × candidate's projected share. Projected share = historical baseline share + swing in percentage points (both are scenario inputs; the baseline itself is Historical data).
- National totals sum all 36 states + FCT. Ranks come from projected national valid votes.
- Geographical spread: a jurisdiction counts toward the spread when a candidate's projected share there is ≥ 25%. Under the configured presidential rule, a candidate must win the highest national votes AND reach 25% in at least 25 of 37 jurisdictions. If nobody meets both, the dashboard shows that a second ballot would be required.
The same deterministic code runs in your browser (live preview) and on the server (saved, auditable runs), so a saved run always matches what you saw. Each saved run snapshots its inputs.
Data integrity rules
- Candidate shares must total 100% in every state/FCT before a projection can be saved. Invalid assumptions are rejected, never silently normalized.
- Normalization is always an explicit analyst action and is reported as such.
- Vote counts use whole-number precision; percentages use high-precision decimals.
- Every import is validated (required columns, numeric ranges, duplicates) and recorded in an audit log.
- Candidate colours are configurable and carry no party endorsement.