Skip to content

devanshsahu47/sales-analytics-pipeline-gcp

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

📈 Automated Sales Data Pipeline & International Growth Dashboard (GCP)

This project showcases a fully automated, end-to-end data pipeline deployed on Google Cloud Platform (GCP). It ingests raw e-commerce sales data, processes it, stores it in a cloud data warehouse, and visualizes key business insights through an interactive Looker Studio dashboard focused on identifying international growth opportunities.


📌 Objective

Build a scalable, serverless data analytics system that provides real-time insights into international sales performance.

Key goals:

  • Automated Data Ingestion: Seamlessly load new sales data into the data warehouse.
  • Exploratory Data Analysis (EDA): Uncover trends and patterns in global sales.
  • Interactive Visualization: Highlight international market performance through a dynamic dashboard.

⚙️ Architecture Overview

The pipeline follows an event-driven, serverless architecture using GCP components:

Stage Tool Purpose
Ingestion Google Cloud Storage Upload .csv files to trigger the pipeline
Processing Cloud Run Serverless Python function processes & loads the data
Storage BigQuery Scalable data warehouse for real-time SQL querying
Visualization Looker Studio Dashboard connects to BigQuery for up-to-date insights

📦 Dataset: E-Commerce Sales Data

Filename: data.csv
A transactional dataset representing sales from an online retail company.

Key Features:

  • InvoiceNo: Unique ID for each transaction
  • StockCode, Description: Product identifiers and names
  • Quantity: Number of units sold
  • InvoiceDate: Timestamp of purchase
  • UnitPrice: Price per item
  • CustomerID: Unique customer identifier
  • Country: Location of purchase

🔍 Key Insights from EDA

  • Market Dominance: The UK contributes ~85% of all sales, indicating high domestic reliance.
  • Order Cancellations: Negative quantities signal frequent cancellations needing further attention.
  • Peak Sales Window: 12 PM to 2 PM emerges as the busiest time for transactions.
  • Guest Checkouts: ~25% of sales lack CustomerID, complicating customer retention analysis.

📊 Dashboard Highlights: International Growth

The Looker Studio dashboard was developed to turn data into strategic direction for global expansion.

Key Metrics & Findings:

  • International Revenue Share: 15.65% – a measurable baseline for non-UK growth
  • Top Non-UK Markets: Netherlands, Ireland (EIRE), Germany, France
  • Higher AOV Internationally:
    • Intl AOV: €521
    • UK AOV: €426
  • Top Intl Products Differ: Best-selling international products differ from domestic bestsellers, influencing region-specific campaigns.

🧠 Tools & Technologies Used

Tool / Tech Purpose
Google Cloud Storage Raw data lake for .csv ingestion
Cloud Run Serverless Python-based data processing
BigQuery Cloud data warehouse for analytical querying
Looker Studio Real-time, interactive dashboarding
Python + pandas Data cleaning and transformation
GitHub Version control and project tracking

🧩 Business Implications

  • International markets, particularly Germany and France, present strong growth potential.
  • Higher AOV internationally justifies investment in:
    • Region-specific marketing
    • Localized websites
    • International shipping capabilities
  • Product marketing should be localized; bestsellers vary by region.

🔗 Live Dashboard

👉 [View the Live International Growth Dashboard] https://lookerstudio.google.com/reporting/27adc2e9-bde6-4e1a-9f30-da73dc777749


📝 License

This project is licensed under the MIT License. Feel free to use, share, and adapt with attribution.

About

End-to-end automated sales data pipeline on Google Cloud Platform (GCP) with BigQuery, Cloud Functions, and Looker Studio dashboard for real-time international growth insights.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors