[ GURU_ANALYTICS_ONLINE ]

MISSION: EXTRACT, ANALYZE, AND VISUALIZE CUSTOMER SENTIMENT FROM UNSTRUCTURED RESTAURANT REVIEWS.

---

// CLOUD_INFRASTRUCTURE //

This project leverages a full Google Cloud Platform (GCP) serverless architecture to ensure scalability and zero-maintenance operations.

[ GCP_SERVICES ]

• Cloud Storage: Data lake for raw JSON reviews and processed datasets.
• Cloud Run: Containerized Python scrapers and NLP microservices.
• Cloud Functions: Event-driven triggers for data processing pipelines.
• BigQuery: Serverless data warehouse for high-speed SQL analytics.

graph LR
    A[SerpApi / Scrapers] -->|Raw Data| B(Cloud Storage)
    B -->|Trigger| C{Cloud Functions}
    C -->|Process & NLP| D[Cloud Run]
    D -->|Structured Data| E[(BigQuery)]
    E -->|Direct Query| F[Power BI Dashboard]

---

// DATA_ENGINEERING_PIPELINE //

1. INGESTION [ WEB_SCRAPING ]

• Integrated SerpApi to fetch real-time reviews from Google Maps and Yelp.
• Custom Python scrapers running on Cloud Run to handle pagination and rate limiting.

2. PROCESSING [ NLP_CORE ]

• Sentiment Analysis: Applied DistilBERT models to classify reviews (Positive, Neutral, Negative).
• Aspect-Based Sentiment Analysis (ABSA): Extracted specific feedback on “Food”, “Service”, and “Ambiance”.

3. VISUALIZATION [ COMMAND_CENTER ]

• Power BI Dashboard: Connected directly to BigQuery.
• KPIs: Net Promoter Score (NPS), Sentiment Trend over Time, Keyword Clouds.
• Impact: Enabled restaurant managers to identify service bottlenecks in real-time.

---

// KEY_METRICS //

• Data Volume: Processed 50,000+ reviews.
• Latency: < 2 minutes from review posting to dashboard update.
• Cost: Optimized serverless usage for < $5/month.