Game Patch Impact & Player Sentiment Intelligence

What this Actor does

This Actor does not simply return Steam reviews. It compares player feedback across two time periods and surfaces sentiment shifts, newly emerging issues, regressions, improvements, and feature requests.

Who it is for

Game studios, publishers, QA teams, community managers, analysts, and game researchers who need a compact update-impact signal from public Steam feedback.

Key features

• Recent-vs-previous, latest-detected-patch, and custom-date comparisons.
• Steam recommendation rates and deterministic sentiment deltas.
• Gaming issue themes, feature requests, short evidence, coverage metadata, and confidence.
• Bounded pagination with partial-coverage warnings and per-game failure isolation.

Supported input

See sample-input.json for the safe Store-test input. The Actor accepts one to ten numeric Steam App IDs, a period length from 1 to 30 days, and a sample cap from 10 to 250 reviews per period.

Output

The default dataset contains one final report per Steam App ID with comparison, impact, patch, newIssues, regressions, improvements, featureRequests, topNegativeThemes, topPositiveThemes, coverage, languageDistribution, warnings, and bounded stats fields. Raw review collections are not published as the primary output.

Output fields

Field	Meaning
status	ok, partial, or failed report status.
steamAppId	Requested numeric Steam App ID.
gameName	Best-effort Steam game name.
requestedComparisonMode	Mode supplied in the input.
effectiveComparisonMode	Mode actually used after patch fallback.
comparison.before.positiveRate	Steam recommendation rate before the boundary.
comparison.after.positiveRate	Steam recommendation rate after the boundary.
comparison.sentimentDelta	After minus before positive recommendation rate.
impact.direction	Thresholded sentiment direction or insufficient_data.
impact.confidence	Bounded 0–1 confidence score.
impact.confidenceLabel	high, medium, or low.
comparison.before.reviewsAnalyzed	Before-period analyzed sample count.
comparison.after.reviewsAnalyzed	After-period analyzed sample count.
coverage.before.coverageStatus	Before-period full, partial, or insufficient.
coverage.after.coverageStatus	After-period full, partial, or insufficient.
newIssues	Complaint themes meeting the new-issue safeguards.
regressions	Themes meeting the regression safeguards.
improvements	Themes with a material negative-mention decrease.
featureRequests	Bounded normalized requests and counts.
warnings	Coverage, fallback, or enrichment warnings.
stats.durationMs	Bounded per-game runtime in milliseconds.
generatedAt	Report generation timestamp.

Limitations

Steam's recent-review API is paginated chronologically. On extremely high-volume games, a run may hit its safety scan limit before reaching the complete requested historical window. In this case the Actor returns partial coverage and lowers confidence instead of presenting incomplete data as a full comparison.

This V1 uses public Steam data and deterministic local analysis. It does not claim that a patch caused an observed change, and it does not collect Reddit, Discord, YouTube, or Metacritic data.

Cost considerations

The recommended pricing principle is pay per successfully generated game report, not per raw review. A starting target is approximately $0.03 per game_report, subject to measured Apify usage economics. This repository does not change pricing automatically.

Compliance and responsible use

Review snippets are short evidence for aggregate analysis, not endorsements. Do not use output to identify, target, or harass individual reviewers. Respect Steam terms and applicable privacy/data-retention requirements.

FAQ

Does this confirm that a patch caused a bug?

No. It reports observational changes in player feedback; engineering teams should reproduce and verify issues independently.

What happens when Steam history is incomplete?

The report marks coverage as partial or insufficient, lowers confidence, and includes a warning rather than claiming full coverage.

API usage

Run the Actor with a JSON object matching sample-input.json using Apify's API or Console. The default input requires no API key, proxy, browser, or external model.

{
    "steamAppIds": ["646570"],
    "comparisonMode": "recent_vs_previous",
    "windowDays": 7,
    "maxReviewsPerPeriod": 40,
    "language": "english",
    "includeOffTopicReviews": false,
    "includeEvidence": true
}

Benchmark results

Phase 0 feasibility measurements are recorded in the repository-level docs/phase-0-feasibility.md. Phase-specific runtime and cloud evidence will be appended to BENCHMARK_NOTES.md as implementation proceeds.