This whitepaper is for historical reference only. Some content might be outdated and some links might not be available.
AWS analytics pipeline and leveraging your data lake
This guide has emphasized moving data in to your data lake that is built on an Amazon S3 bucket. Next, you learn about what you can do with the data you have collected. Archival data can be an invaluable source when performing analytics and model training for AI/ML. The archival data provides historical comparison to accurately determine trends, and provide helpful insights. For more information, see the AI/ML for workflow optimization section.
Analytics pipeline and leveraging your data lake
Starting with all of the data that you have collected, both current and archival, you use
several AWS managed services, to include visualizing your data with an Amazon QuickSight
QuickSight dashboard showing wafer defect types and foundry location
In this example, the dashboard shows the defect type and the location of the fab where the defect occurred. From this data, it appears that defects are occurring about the same rate and type across all of the fabs. If, for example, there was an increase in defects at a specific fab when fabrication started on your design, this may indicate a problem with the PDK or another design issue. With this insight, you can make decisions about designs in-flight, that could result in reduced re-spins and increased fabrication yields.
