The rapid growth of retail data presents significant challenges for organizations seeking to transform raw information into actionable business insights.

Introduction

Retail organizations generate large volumes of data across multiple business domains, including customer information, product catalogs, sales transactions, and customer feedback. As data volumes continue to grow, organizations require scalable platforms capable of storing, integrating, and analyzing data efficiently.

Traditional approaches often rely on manual data movement and fragmented analytical processes, making it difficult to maintain data consistency and derive timely business insights. A cloud-native analytics architecture can address these challenges by automating data ingestion and providing a centralized platform for large-scale analytical processing.

This article demonstrates how Alibaba Cloud Object Storage Service (OSS), DataWorks, and MaxCompute can be combined to create an end-to-end retail analytics pipeline. The solution enables organizations to ingest raw retail datasets, automate data integration workflows, and prepare data for analytical processing within a scalable cloud data warehouse.

Solution Overview

The proposed solution uses OSS as the centralized storage layer for raw retail datasets, DataWorks as the data integration and orchestration platform, and MaxCompute as the analytical data warehouse.

Retail datasets are first uploaded to OSS and organized according to business domains. DataWorks Data Integration is then used to synchronize the datasets into MaxCompute tables, where the data becomes available for analytical queries and reporting workloads.

This architecture separates storage, integration, and analytical processing responsibilities while providing a scalable foundation for enterprise data platforms.

Architecture Design

diagram_june
The architecture consists of four primary layers.

Data Storage Layer

OSS serves as the centralized storage repository for raw retail datasets. Data files are uploaded in CSV format and organized according to business domains.

Data Integration Layer

DataWorks automates the ingestion process by reading datasets from OSS and synchronizing them into MaxCompute tables. This eliminates manual data loading processes and improves operational efficiency.

Data Warehouse Layer

MaxCompute provides a fully managed and highly scalable data warehouse capable of processing large analytical workloads without requiring infrastructure management.

Analytics Layer

Business users and data analysts can execute SQL queries directly against MaxCompute to generate reports, identify trends, and support data-driven decision-making.

Preparing Retail Datasets in OSS

To simulate a retail analytics environment, multiple datasets were prepared and uploaded to OSS.

The datasets were organized into separate folders representing different business domains.

raw-data/
├── customers/
│   └── customers.csv
├── orders/
│   └── orders.csv
├── products/
│   └── products.csv
└── reviews/
    └── reviews.csv

Retail Dataset Structure

The customer dataset contains customer profile information, while the order dataset contains transactional records used for analytical processing.

By storing raw files in OSS, organizations can establish a centralized data lake layer that supports downstream analytics and data warehousing workloads.

Creating Analytical Tables in MaxCompute

Before loading the datasets, destination tables were created within MaxCompute.
Screenshot_2026_06_13_at_06_53_44
The order table was defined as follows:

CREATE TABLE retail_orders (
    order_id STRING,
    customer_id STRING,
    product_id STRING,
    order_amount DOUBLE,
    order_date STRING
);

MaxCompute Table Creation

These tables serve as the analytical foundation for the retail data warehouse.

Building the Data Integration Pipeline

Alibaba Cloud DataWorks Data Integration was used to automate the movement of data from OSS into MaxCompute.
Screenshot_2026_06_13_at_06_51_03
A synchronization task was created using OSS as the source and MaxCompute as the destination.

Source Configuration

The source dataset was configured using the OSS data source.
Screenshot_2026_06_13_at_07_02_46

Source Type: OSS
File Format: CSV
File Path:
raw-data/customers/customers.csv

Destination Configuration

The destination was configured using the MaxCompute data source.
Screenshot_2026_06_13_at_07_02_51

Destination Type: MaxCompute
Project: retail_analytics_project
Table: retail_customers

Executing the Data Synchronization Job

After configuring the synchronization task, the pipeline was executed through DataWorks.
During execution, DataWorks performed the following operations:

Read the source CSV file from OSS.
Parsed and validated the incoming records.
Mapped source fields to the destination schema.
Loaded records into MaxCompute.
Generated execution logs and monitoring information.

The execution completed successfully and processed all records contained within the dataset.

Successful Data Synchronization

Example execution summary:
Screenshot_2026_06_13_at_07_04_16

Job completed successfully.
Total Records Processed: 20

The successful execution confirms that the retail dataset was ingested from OSS into MaxCompute through DataWorks Data Integration.

Exploring Data in MaxCompute

Once the synchronization process completed, the datasets became available for analytical workloads within MaxCompute.
Screenshot_2026_06_13_at_06_54_00
The imported data can be queried using standard SQL statements.
Example query:

SELECT COUNT(*)
FROM retail_customers;

Additional analytical queries can be performed to explore customer distributions, purchasing behavior, and sales trends.

Data Exploration in MaxCompute

This capability enables organizations to transform raw operational data into actionable business intelligence.
Screenshot_2026_06_13_at_07_10_07
To demonstrate the analytical capabilities of MaxCompute, a product sales analysis was performed to identify the best-selling products based on total quantities sold. This query aggregates order quantities by product and ranks products according to their sales volume, enabling retailers to better understand customer demand patterns, evaluate product performance, and support inventory planning decisions.

Benefits of the Architecture

Scalable Analytics

MaxCompute provides a serverless analytical platform capable of processing large-scale datasets without infrastructure management.

Centralized Data Storage

OSS serves as a durable and cost-effective repository for raw retail data.

Automated Data Integration

DataWorks simplifies data ingestion workflows through visual configuration and centralized orchestration.

Cost Efficiency

Organizations can separate storage and compute resources while scaling analytical workloads according to business requirements.

Enterprise Readiness

The architecture supports data governance, operational monitoring, and future expansion into advanced analytics and machine learning workloads.

Conclusion

Alibaba Cloud OSS, DataWorks, and MaxCompute provide a powerful combination for building modern data analytics platforms.

By storing raw datasets in OSS, automating ingestion through DataWorks, and leveraging MaxCompute for large-scale analytical processing, organizations can establish a scalable and efficient retail analytics architecture.

This implementation demonstrates how retail data can be transformed into actionable insights through a cloud-native data platform while reducing operational complexity and enabling future analytical initiatives.

Community

Transforming Retail Data into Actionable Insights with OSS and MaxCompute