Tech Eval 2019-08-13 Meeting notes

• What can we achieve in the upcoming weeks:
• Start to categorize potential approaches that the group can review, comment and add pros and cons to. Would like to have a first draft prior to the Summit so we can use these as part of our working session discussion.

Self-Contained:

 Data Flow - Into the ODS and Out of the ODS

               Data Store - ODS

               Front End - Connects directly to ODS for queries

Self-Contained++:

               Data Flow - Into the ODS and then Into the Analytic Middle Tier

               Data Store - Separate Schema in ODS

               Front End - Connects to Analytics Middle Tier and/or ODS

All External:

               Data Flow - Into the ODS and then Out of the ODS to external system (PostgreSQL, Big Query, etc...)

               Data Store - Separate stand alone database

               Front End - Connects to the external source only

Data Warehouse/Lake:

               Data Flow - Into the ODS and then into read-optimized data architecture

               Data Store - Distinct schema but database can be on the same or external server to ODS

               Front End - Connects to the data warehouse/lake infrastructure

So far approached reviewed do not query the ODS directly, they have added/developed an architecture that moves data out and manages/manipulates for reporting.

• • If we believe that it is best to query a data store, should we develop a standard for analytics tier so that views are the same regardless of the technology used. This will allow the community to then develop connectors or mappings between the analytics core views and whatever they might be using.
    
    • Initial set of views used by innovateEDU https://docs.google.com/spreadsheets/d/10EhwdMyb015thaPAXbGFWm0Pp_CCATmMiyL2nUeNWiw/edit#gid=1347860255.
    • These have evolved based on user feedback, would be great to collaborate with the community on improving or adding onto these.
  • Could we get more info on the data warehouse project.
  • Start capturing views used in the field as analytics middle tier.