Modernizing Data Lake & BI Platform For a Can Manufacturing Major
ManufacturingBusiness Impacts
Improved business efficiency by 80%
Reduction in cost by means of Pay-per-use pricing and serverless architecture
Improved process efficiency by reducing delays in data processing, ETL jobs, and reporting
- Location : USA
- Industry : Manufacturing
Problem Context
The client is the world’s leading supplier of two-piece can and end-making machinery for the global can-making industry. They supply individual machines, and provide design, installation and support for complete can and end lines for beverage and food cans.
The client’s business team relied on ad-hoc reports to make business decisions. They needed an enterprise-scale data analytics and BI reporting platform.
Challenges
- Creating a unified Data Model with different datasets
- Integrating data from various sources (stream and batch) of different sizes and data structures
- Several hierarchies (Tier 3, 4, 5) of dashboards to be built for other users and personas
Tools & Methods
Amazon QuickSight
Amazon Redshift
Amazon S3
Amazon Lambda
AWS IAM
AWS CloudFormation
Amazon Athena
Amazon CloudWatch
AWS Glue
Amazon VPC
AWS Glue Data Catalog
AWS Glue Crawler
AWS Secret Manager
Amazon Kinesis Data Firehose
AWS SNS
Solution
Quantiphi designed and implemented the AWS Lake House architecture with best practices. Collecting and segregating data at different levels and end-to-end BI platform development enables better insights into their factory performance.
Results
- Processed 200 records per minute from factory sensors installed in two factories in the UK and Hungary
- A Data Lake House architecture was created from streaming and batch data from different sources and factories
- Three Quicksight Dashboards were created, covering 10 to 12 KPIs relevant to distinct users
- The dashboards were refreshed with a latency of 10-15 minutes with near real-time data
