Section-level risk screening

Each polling section receives a composite 0–100 score combining six independent statistical signals already published as standalone reports. A high score means the section is statistically unusual along multiple dimensions and warrants a closer look — that is all.

The six signals

  • Recount delta — discrepancies between the first and second tally of ballots.
  • Flash-memory mismatch — gap between the official protocol and the machine-voting device's protocol.
  • Invalid-ballot share — sections with an unusually high share of invalid ballots.
  • Additional voters — sections with a disproportionately high count of voters added on election day.
  • Vote concentration — sections with a suspiciously high single-party share.
  • Peer-section outliers — results that diverge statistically from nearby sections in the same settlement.

See the full methodology for thresholds, formulas, and bands.

Explore the data

  • Home
  • Governance
  • National Assembly
  • Roll-call votes
  • Governments
  • MP business connections
  • Public procurement
  • State budget
  • EU funds
  • Indicators
  • Consumption
  • Prices
  • Party financing
  • Opinion polls
  • Seat simulator
  • Parties
  • Results by region
  • Extraordinary local elections
  • Local-elections reconciliation
  • Sofia
  • About the project

Recent analysis

  • Ask Naiasno: a new way to explore Bulgaria’s public data
  • The money map: where the state buys and where the money goes
  • We are opening up the procurement risk flags