Continuing the discussion from 2018 Collaborative Data Innovations for Sustainable Development - Call for Collaboration on OpenMRS Proposal:
In the end we will not submit for this World Bank funding opportunity, but this topic recently came up, I would like to explore a bit further whether this tool could be used as the basis for a standard OpenMRS Data Warehouse solution.
Basically, a Bahmni project team has done some nice work in creating a reusable tool that:
- Flattens/simplifies the OpenMRS data model into an analytics-friendly format
- Is configured via JSON files that indicate which transforms you want to do
- (optionally) puts metabase on top of this, so you can immediately do analytics
- More details in this thread: Creating a Bahmni Mart database for analytics
Under the hood it uses Spring Batch, and stores data in postgres.
In theory this should work* out of the box for a plain OpenMRS implementation. Would anyone be interested to test this out, and share some feedback on whether it seems like a worthwhile direction for a generic/standard OpenMRS tool.
- = one thing that will not work out of the box for plain OpenMRS is the ability to automatically flatten to row-per-encounter & column-per-question, because this is driven by the Bahmni form-entry technology. One could add similar code to support HTML Form Entry, XForms, etc, if this seems like a promising direction.