Methodology

1. What we ask

Every rating has six 1–5 questions:

• 🛣️Roads & cleanlinessweight 15%
• 🏛️Govt office serviceweight 20%
• 🧾Bribe-free experienceweight 25%
• 🚦Traffic & transportweight 10%
• 🛡️Safety & lawweight 10%
• ✨Hope in governmentweight 20%

Weights reflect what citizens told us matters most day-to-day: bribery (25%) and hope (20%) carry the highest signal.

2. Score formula

Each 1–5 rating becomes 0–100 via (r − 1) × 25. Area score is a trust-weighted average across all responses.

3. Trust weights

• • Anonymous rating → weight 0.5
• • Anonymous + office visit context → weight 0.8
• • Verified user (coming soon) → weight 1.0
• • Suspicious/flagged → weight 0 (excluded until reviewed)

4. Confidence labels

• • High: 50+ ratings and ≥30% verified
• • Medium: 20+ ratings
• • Low: 5–19 ratings
• • Insufficient: < 5 ratings — shown privately but not ranked publicly

5. Anti-spam

We auto-flag ratings when the same browser submits to the same area again within 30 days, when an IP submits more than 5 ratings an hour, when identical comments recur, or when a response is all 1s or all 5s without any context. Flagged ratings don't contribute to scores until reviewed.

6. Limitations

This is citizen perception, not measurement. Areas with fewer responses are less reliable. Ratings reflect who chose to respond — not a random sample of the population. Read the numbers with that in mind.