/

Case Studies

/

Intelligent order optimization

Intelligent order optimization

Intelligent order optimization

How Databricks enabled scalable, efficient data processing across regions

How Databricks enabled scalable, efficient data processing across regions

Seamless migration

Efficient version control

Automated pipelines

Global scalability

The challenge

Navigating scalability and localization across countries

Fractal implemented an advanced smart-order solution using R Studio at the store level, designed to support operations across three regions in Pakistan. This solution was then successfully scaled to cover Pakistan, India, and Indonesia, ensuring simultaneous output generation across all regions.

Key challenges

  • Customizing for regional needs

  • Adapting the solution across regions

  • Ensuring timely generation for all regions

The solution

Scalable and efficient data management

Scalability and integration

Migrated to Databricks for scalability

Integrated parallel pipelines across regions

Used PySpark for SQL Server data extraction

Efficient data management

Centralized storage in Blob

Optimized computing with DBFS

Enabled smooth file handling via DBFS

Implementation approach

1

Databricks setup

  • Migrated to Databricks for cloud execution

  • Optimized performance and scalability

  • Configured pipelines for parallel processing across regions

2

Data flow management

  • Extracted data with PySpark from SQL Server

  • Stored and ensured consistency in Blob using DBFS

  • Enabled real-time processing

3

Script automation

  • Managed child snippets with Master Script

  • Automated execution with parameters

  • Simplified code management

The impact

Scalable solutions with optimized code and data flow

Scalability

  • Enabled parallel execution across regions

  • Migrated to production for better scalability

  • Automated pipelines using Azure Data Factory

Code management

  • Integrated GitHub for version control

  • Simplified deployments and updates

  • Improved collaboration

Data processing

  • Enabled real-time R execution

  • Streamlined workflows

  • Reduced manual effort

Looking ahead

Scalability

  • Expand to new regions and handle more data

Automation

  • Boost efficiency with enhanced automation

Optimization

  • Improve processing speed and real-time analytics