Google Play - Reviews Sentiment Analysis

Google Play – Reviews Sentiment Analysis

This actor analyzes hundreds or thousands of reviews of an Android app on Google Play Store in a few minutes and generates a ready‑to‑use strategic report, powered by advanced LLM models running on Groq (including GPT‑OSS 120B for the Premium tier).

What the actor does

• Automatically extracts an app's reviews from the Google Play Store using only the package name (appId).
• Processes reviews in batches, extracting recurring themes organized into pros (strengths) and cons (issues).
• Aggregates themes into synthetic statistics and generates a structured report with an executive summary, strengths, weaknesses, and actionable recommendations.

External Actors and billing

• This Actor internally calls the public Actor curious_coder/google-play-scraper to collect Google Play reviews.
• All costs related to the usage of curious_coder/google-play-scraper (including any rentals or pay-per-event fees) are fully included and covered in the price of this Actor for the end user.
• End users do not need to pay separately for curious_coder/google-play-scraper; running this Actor is sufficient to cover all underlying Actor costs.

Tiers and LLM models

Tier	Max reviews (hard cap)	Analysis model (map)	Report model (reduce)	Main use case
Base	~200	llama-3.1-8b-instant	llama-3.1-8b-instant	Balanced analysis for ASO, product and UX.
Complete	~500	llama-3.1-8b-instant	llama-3.1-8b-instant	Balanced analysis for ASO, product and UX.
Premium	~2,000 (also limited by maxReviewsToProcess)	llama-3.1-8b-instant	gpt-oss-120b	Highest‑quality reports with very detailed summaries.

The actual number of processed reviews is the minimum of maxReviewsToProcess, the tier limit, and 2000.

PPE billing (Pay‑per‑event)

Billing is handled via a single PPE event per run, with a different event for each tier.

Selected tier	PPE event called	Billing frequency
Base	BaseAnalysisCompleted	1 event per completed run.
Complete	CompleteAnalysisCompleted	1 event per completed run.
Premium	PremiumAnalysisCompleted	1 event per completed run.

The actual cost per run is defined in the PPE configuration of your Apify account for each event.

Input parameters

Field	Type	Required	Default	Short description
appId	string	Yes	–	App package name on Google Play, e.g. com.fake.AppSentimentAnalysis.
pricingTier	string	Yes	complete	base, complete, premium; controls review limits and models used.
detailLevel	string	No	low	low, medium, high; controls report depth and length.
maxReviewsToProcess	integer	No	500	Upper limit of reviews to download and analyze for the run.
summaryPrompt	string	No	Generic default prompt	Optional prompt to adapt the focus (ASO, UX, marketing, etc.).
recentDays	integer	No	0	Only include reviews from the last N days (0 = no limit)

Output structure

The actor saves a single summary record in the default run dataset.

Field	Description
appId	Package name of the analyzed app.
pricingTier	Tier used for the analysis (base, complete, premium).
detailLevel	Detail level chosen for the report.
summaryReport	Full AI report text with executive summary and recommendations.
topPros	List of most frequent positive themes with counts.
topCons	List of most frequent negative themes with counts.
totalReviewsProcessed	Total number of reviews actually processed.

Example Output (Complete JSON)

{
  "app_id": "com.fake.AppSentimentAnalysis",
  "pricing_tier": "complete",
  "detail_level": "medium",
  "summary_report": "### Executive Summary\nFake Sentiment Analysis app receives generally positive reviews (4.2/5 average) with praise for intuitive interface and analysis speed. Main criticisms concern free plan limitations and instability when exporting large reports.\n\n### Top Strengths\n1. **User-friendly interface** (42 mentions): 'Clean and easy-to-use dashboard'\n2. **Fast analysis** (35 mentions): 'Results in just a few seconds'\n3. **Useful insights** (20 mentions): 'In-depth sentiment analysis'\n\n### Top Weaknesses\n1. **Limited free features** (18 mentions): 'Basic functionality only in free plan'\n2. **Export bugs** (10 mentions): 'Issues with large CSV files'\n3. **Learning curve** (8 mentions): 'Onboarding could be improved'\n\n### Recommendations\n- Improve export stability for reports >1000 reviews\n- Clarify plan limits on landing page\n- Add interactive tutorial for new users\n- Offer 7-day Premium trial instead of 3 days",
  "top_pros": [
    [42, "User-friendly interface"],
    [35, "Fast analysis"],
    [20, "Useful insights"],
    [15, "Clean dashboard"],
    [12, "Real-time results"],
    [8, "Mobile responsive"],
    [6, "Export options"],
    [4, "Dark mode"],
    [3, "API integration"],
    [2, "Multi-language"]
  ],
  "top_cons": [
    [18, "Limited free features"],
    [10, "Export bugs"],
    [8, "Learning curve"],
    [7, "Slow on mobile"],
    [5, "No bulk upload"],
    [4, "Premium expensive"],
    [3, "No templates"],
    [2, "Poor support"],
    [1, "No offline mode"],
    [1, "Ads intrusive"]
  ],
  "total_reviews_processed": 247
}