VoteWatch NG

Let any Nigerian citizen, accredited observer or journalist report electoral misconduct from a polling unit as it happens, and produce evidence that still means something months later. Existing tools collect reports. None of them produce anything a tribunal will accept.

A photograph of a ballot box being carried away is worth very little on its own. A tribunal needs to know when it was taken, where, by whom, and that nobody altered it in between. Reporting tools almost never carry that, so the evidence dies somewhere between the polling unit and the hearing. So the product was designed backwards from the tribunal rather than forwards from the report form. Media is hashed with SHA-256 at the moment of capture, with location and timestamp injected before the file is ever written, and every status change appends to an audit log instead of overwriting one. The second constraint is where this actually gets used: rural polling units with almost no signal, on the one day of the year when every observer in the country is submitting at once. Capture works offline and the queue drains when a connection returns.

• Twelve misconduct categories drawn from patterns documented across four general elections
• SHA-256 hashing at capture, with location and timestamp injected before the file is written
• Chain of custody: every status change appends to an audit log rather than overwriting one
• Tribunal-ready evidence bundles exported as PDF or ZIP for legal escalation
• Real-time incident map with category pins and clustering, to expose density and pattern
• Results verification against official polling unit figures, flagging discrepancies
• Anonymous reporting and an SOS path, because observers are routinely intimidated
• Offline-first capture for low-connectivity polling units, syncing when signal returns

• Flutter
• Dart
• Firebase
• Firestore
• Google Maps

• 12Misconduct categories
• SHA-256Evidence chain
• OfflineFirst capture
• 2025In build