I’m looking for some advice from those experienced/ knowledgeable in dashboard development or content management
Here’s the story:
We’ve now released v2.0 of the Measures Library, and we’re rather chuffed with it It’s published using r-shiny, has 188 measures, a bunch of new documents and features, etc etc. However, we’d like to get more efficient with our systems for production to help scale it further, faster.
Here’s the rub:
The Measures Library has considerably more text content, with more frequent changes, than what our team typically deals with. A lot of this content is published as PDF documents as well as being used in the app. At the moment we are manually maintaining two versions of this content – one for PDF and one for Shiny. This is becoming increasingly unwieldy, and we are looking for solutions that will:
a) Let us have a single source for both the app content and the PDFs
b) Let non-technical people write, edit, comment and sign off the content.
You can find more detail what I’m talking about re: content management challenges in this article, if you’re interested
Our ideas:
Store the content in a database. Connect the app for the app content. Pull into RMarkdown or similar for the PDFs (this only meets requirement a))
Host the content on an off-the-shelf CMS with the kinds of features we want for requirement b). Still use Shiny for the data visualisation by embedding the app in an iframe on this new site. We would need to connect the app and CMS such that when the user selects a different category / measure on the app, the correct content is displayed in the CMS. We are not sure how to do this or whether it’s worth the effort. Also not sure what the backend of such a CMS is like, but presumably there is a database somewhere which we could still use for RMarkdown as in 1.
The limits:
We don’t feel particularly inclined towards moving analytics off r-shiny (Posit) to something like tableau, power BI or Qlik. They have their own set of issues, and we’re not convinced that they will inherently solve this conundrum anyhow. That said, if you think they would, then please do convince me
If you have some ideas about how one might go about solving these issues, or even if you just know a person/ org that can, I’d love to hear from you!
It would also be worth adding links to the Posit interactive dashboarding product Shiney. Also, they have released their next-generation documentation platform after RMarkDown, Quarto, which I am currently slowly exploring with positive results to date.
I think there are two parts here which people often don’t consider as separate activities. The first part is the process for generating and publishing the content (a workflow with approval/release gates). The second part is publishing the content which might be separate or might be combined.
I can suggest two people to talk to about this:
@Tracey_Vandenberg - her team publish the national collections and they have processes for both publishing and workflow
@Merrin_Macleod - Merrin is working on project buttercup which is looking at our content publishing approach in Te Whatu Ora (in part driven by the 100+ websites we have) to reduce duplication and improve the quality of the publishing we do. The focus is communications for consumers.
We are working on something similar in terms of the Medicines Information Database. We are using Jira Service Management for the workflow with Confluence to store/manage the content.
Interested to hear more about your experience with Quarto - what are you using it for currently?
Thanks for the tips Jon - will hit up Tracey and Merrin - figured someone out there must have cracked this! Have been poking around a few other agencies as well so will be interesting to see what falls out
Keen also to learn more about how you’re using Jira and confluence - will DM you
I’m also keen to learn more about them and how they fit in with Te Whatu Ora / the wider NZ digital health ecosystem. Perhaps someone is willing to give us a demonstration and answer some awkward questions?
It’s published using r-shiny, has 188 measures, a bunch of new documents and features, etc etc. However, we’d like to get more efficient with our systems for production to help scale it further, faster.