At this year’s CTEC, a few colleagues and I presented plenary sessions on what LABVANTAGE was doing in the areas of data services and business intelligence. One of the items we presented was a business intelligence solution geared toward laboratory informatics called “DecisionScapes.” The goal of this solution was to create a highly valuable business intelligence and data warehousing solution that provided out-of-the box key performance indicators (KPIs) in support of several laboratory informatics domains. As we have put it before, a zero time-to-value business intelligence solution for laboratories.
With that goal in mind, I have drafted the first building block of this concept, which is a data warehousing data model geared toward QA/QC for Manufacturing. This data warehouse data model is designed around several KPIs that were identified as being critical to a successful QA/QC manufacturing environment.
Here are a few example KPIs for which the data model is designed:
Executive Level KPIs
- Supplier Performance
- Customer Complaints
- Cost Per Analysis
- Cost Per Product
Laboratory Manager KPIs
- Defects in Final Products
- Equipment Quality
- Product Released on Target Time
- Retests Per Analyst
We have identified a total of 18 KPIs as being critical to making informed decisions in the QA/QC manufacturing environment, and the data model has been constructed to support all of them.
The data model is structured to use standard data warehousing star and snowflake schema designs, which means that the structure of the warehouse will be compatible with how modern ETL and visual intelligence software packages like to operate. In fact, if you attended either the business intelligence training class or the data integration training class on Thursday during CTEC 2011, you already used an early draft of this data warehouse data model using the expressor ETL and Tableau business intelligence software packages.
On top of the standard dimension, fact, bridge, and flag database tables, the data model comes with a slew of data views that have business user-friendly names. No underscores, no acronyms, just clearly identifyable business names for common QA/QC manufacturing activities and objects. In addition to including user-friendly views on all base tables, the data model comes with similarly user-friendly views that are based on common table binds. Do you need to see all results for a particular material? There’s a view for that! Do you need to see all test for a particular batch? There’s a view for that!
So, what comes next?
Obviously, we will eventually need to create the ETL and visual intelligence layers below and above this data warehouse data model. But before that can happen, we need to partner with some QA/QC manufacturing organizations that would be interested in testing out this data model and providing us feedback on how it is being used and to help us identify the inherent strengths and weaknesses of the model.
Interested in participating? Contact us! Feel free to ask for me by name, and we can work together on creating a QA/QC for Manufacturing data warehouse that collects and aggregates data in a way that helps you make daily decisions with confidence.