Data-driven game review analysis

Game Review Analysis

We process millions of Steam review events to surface the patterns, anomalies, and trends that matter to gamers, developers, and industry observers.

Reviews tracked

160M+

Games monitored

10,000+

Snapshot cadence

3 min

How the analysis works

Review Velocity

We measure how many reviews arrive per hour and compare that rate against each game's baseline. A 10x spike in a single hour signals anomalous behaviour โ€” whether positive or negative.

Sentiment Scoring

Every snapshot captures the ratio of positive to negative reviews. We track how that score shifts over time relative to the long-run average โ€” identifying recoveries, declines, and inflection points.

Tier Classification

Incidents are classified as WARNING โ†’ CRITICAL โ†’ NUCLEAR based on both burst velocity and sustained pressure. Classifications update in real time as more data arrives.

Recovery Tracking

After an incident ends, we continue monitoring the recovery arc โ€” how quickly the score returns toward baseline, and whether the community forgave or forgot.

Cause Tagging

Incidents are labelled with cause tags: anti-cheat rollouts, pricing changes, DRM controversies, patch backlash, and more. This enables cross-industry pattern detection.

Franchise Rollups

Track incident patterns across entire franchises โ€” Elder Scrolls, Dark Souls, Call of Duty โ€” to identify publisher and studio-level behavioural trends.

Recent analysis

All analysis