Internationalization Work Group 2019-12-06

Materials

Previous Action Items

Gene and Linda plan to have an update for the group around rostering and education organizations at the next meteing.
Ed sent out a slide deck to discuss some of the history of Ed-Fi, key structure, abstraction, and referential integrity, and data quality

Presentation of "Historical" Slide Deck

Lots of discussion around the trade off between data quality and the ease of use for a model. No easy answer here - so this will need to be considered throughout the internationalization discussions.
Suggestion that we try to define specific use cases we want to solve and have those be the driving force behind the trade off.
To this end, there may be different use cases depending on the type of Education Organization. This was also mentioned in the previous meeting to possibly categorize Education Organizations by primary function (i.e., enrollment vs. administration).
Another distinction around data quality that came up is who is responsible? Should "dirty" data be allowed if it provides a path to other data points that could prove useful? Perhaps employ warnings instead of hard omissions.
Overall, this balance of how to continue to uphold Ed-Fi design ideas while also allowing for a more flexible, and easy-to-use model will be key to acceptance.

New Action Items

Eric would like to look at the FHIR health models that have scaled internationally. May only have access to the spec publicly but this would provide a good idea. Craig has asked to join that effort. Craig has background in creating data lakes and BI environments in the insurance industry which deals with how to handle incoming "dirty" data.

New Topics to Cover

Could we include metadata around the quality of incoming data? Where the API captures if data is "dirty" or "clean, if dirty then in what way, and any steps taken to fix the data?