We’ve had a number of previous posts on how we can improve analytics: Advanced Reporting/Analytics for OpenMRS, an out of the box ETL solution for OpenMRS, Analytics on FHIR, an ELT/ETL Pipeline Demo, and the previous ETL/ELT standup.
The core of the matter is we want to improve the tools for reporting and analytics on the OpenMRS platform - and we want to be clear on the prominent use-cases faced by the community, so we can start together by focusing on examples that will offer high value for multiple groups.
We’re gathering requirements as we will soon open a squad and tackle the technical next steps (previously referred to as the “ETL Squad”). @akimaina, @bashir, and @aojwang are the technical leads and we’re keen to hear your responses here. Concrete examples of common problems will help focus the new squad’s efforts from the start.
Right now, there are 2 things you can do before we kick off the squad:
We want to hear your use cases - and the problems you’ve faced
What do you want to do with your data?
What makes it hard for you to do that right now?
Vote on the examples that resonate with you
For example, we’ve heard pain points like these ones below - vote on the ones that resonate with you/your implementation!
- Storage complexity makes it difficult to query data quickly
- Not being familiar with the OpenMRS data model makes it difficult to query data
- Big data stores are difficult to manage once beyond the query ability of simpler tools like MySQL
- Aggregating data in your data warehouse is painful
- Maintenance is painful including indicators/queries definition, pipeline maintenance, and servicing Ad Hoc requests
- Risk of impact on operational/production database performance while performing analytics
- Unforgiving analytics/unfriendly environment (e.g. having to rebuild entire analytics pipeline due to new forms or concepts, or a misdefined indicator)
- Other (please share in the comments below)
If you’re using OpenMRS and want improved reporting and analytics tools, simply reply to this Talk post, and/or vote on the examples above with what resonates!
@akimaina @bashir @aojwang @mksd @mseaton @mogoodrich @jdick @nkimaina @paulamendola @burke @ibacher @dkayiwa @ssmusoke @morrisng @ggomez @mozzy @k.joseph @chris @ccwhite23 @janflowers @ayesh @rubailly @danfuterman @wyclif (If I’ve neglected to tag anyone, please do add them below )