Lam Research Classification Data Migration

Lam Research Corporation engineers, manufactures, and maintains semiconductor processing equipment used in the creation of integrated circuits.

Overview

To setup our classification system, we needed to import all of the comercial off the shelf parts. I was in charge of moving 2/3rds of the parts into the system.

Challenges

• Two different vendors were used to gather the classification data which each used different formats.
• Some of the data was retrieved directly from the manufacturer, which had a unique format.
• Lots of the data was dirty after being harvested. We had instances duplicate values, for example SST, SS, Stainless, Stainless Steel all being used to identify stainless steel.
• The team wanted the system to be available as soon as a minimum threshold of classes was imported.

Solutions

• I reviewed all of the data being migrated into the system and cleaned up any areas that needed to be normalized.
• I developed the project plan and managed the 3 member team and migration of 80,000 parts.

Results

• We imported over 102,000 commercial off the shelf parts into the system.
• Our final taxonomy had 483 searchable classes.
• Users were able to start finding parts a year before the entire system was imported.