Methodology

Technical details explaining how PTBI categorizes information, produces analytics, and renders geographic visualizations.

calendar_todayLast updated: August 22, 2026

Contents

info

Please review this document carefully.

This document is provided for information about platform policies and should not be treated as legal advice. Reports on PTBI are user-submitted allegations unless independently verified.

1. Categorization Hierarchy

Every report submitted to PTBI follows a strict data hierarchy to ensure filterability across the platform:

Experience (e.g., Report)
├── Type (Corruption vs Honesty)
├── Target (Government vs Private)
├── Institution / Department (e.g., Police, Bank)
├── Category (e.g., Bribe Demanded, Unofficial Fee)
├── Location (State → District → City)
├── Time (Date of occurrence)
├── Evidence (Has attachments / none)
├── Moderation Status (Pending / Published / Rejected)
└── Verification Status (Unverified / Verified)

2. Map Visualization Methodology

The Map view aggregates reports at the state or district level using geospatial grouped queries. Map color intensity corresponds to reporting volume or the ratio of corruption to honesty reports for a given institution.

Map visualizations represent submitted reporting activity rather than objective corruption levels. Blank areas on the map indicate a lack of platform adoption in that region, not an absence of corruption.

3. Analytics Methodology

The Analytics dashboard relies on dynamic SQL aggregations (via Prisma ORM) against the published reports database.

  • Time Series: Plots report counts aggregated by day, week, or month.
  • Institutional Breakdown: Calculates the percentage share of reports targeting specific departments within the current active filter context.
  • Financial Impact: Sums the user-declared amount field. These totals are self-reported and unverified.

4. Bias and Interpretation Warning

warning Interpret before comparing

Raw counts on PTBI can be highly misleading if compared directly without accounting for structural biases.

  • Sample Bias: The data represents only citizens with internet access who chose to use PTBI.
  • Reporting Bias: People are generally more likely to report negative experiences (corruption) than positive ones (honesty).
  • Population Differences: A state with 100 million people will naturally generate more reports than a state with 10 million.
  • Duplicate Submissions: Highly publicized local events may trigger multiple users to submit reports about the exact same incident.

Related

Data & TransparencyReportsMapAnalytics