The mood index
Each vote is one of five faces, scored 1 to 5. We linearly rescale to a 0–100 scale using (score − 1) × 25, so face 1 maps to 0 and face 5 maps to 100. The published Mood Index for any bucket is the arithmetic mean of those rescaled scores across qualifying entries for the current UTC day.
Suppression threshold (k-anonymity)
Any published number — global, country, segment, or trend point — requires at least 50 contributors in that bucket. Below the threshold, the value is suppressed in dashboards, tooltips, lists, comparisons, the 7-day trend, and any export.
Fixed publication windows
Public aggregates are not served live. They are built from a snapshot that rebuilds every 15 minutes. The snapshot timestamp is shown in the dashboard header. This defends against near-threshold differencing — a vote cast a minute ago does not move the public number until the next snapshot.
Segmentation
Segment comparisons are restricted to a whitelist: device type, language, country, age range, gender, industry, employment status, and education. Arbitrary field probing is rejected server-side. Sensitive attributes (employer, politics, religion, ethnicity, sexuality) are opt-in only and never appear in public aggregates.
Abuse handling
Votes flagged by our abuse heuristics (rapid retries, identity rotation, etc.) are excluded from every aggregate, snapshot and export. Flagging persists across refreshes. Read endpoints are rate-limited per identity to throttle repeated probing.
Limitations
- Self-selection bias. Participants opt in. The sample is not a probability sample of any country or population.
- Change over absolute level. Day-to-day movement is more reliable than comparing absolute levels between countries.
- Coverage. Country and segment indices reflect who chose to vote, not who lives there.
- Residual inference risk. Sophisticated longitudinal correlation across snapshots can still produce weak inferences in extreme cases.
Citation
Cite as: MoodPulse — Global Mood Index, accessed [date]. Source: https://app.moodpulse.today. If you need an archival snapshot for publication, see for media or for research.