Back in December I worked with a colleague from the Azure team, Kunal Jain, to release an end-to-end Azure Synapse and Power BI solution using real CMS Medicare Part D data. The solution can be deployed with a few clicks in Azure and runs in less than an hour. Here is a link to the December article: https://techcommunity.microsoft.com/t5/healthcare-and-life-sciences/deploy-an-end-to-end-azure-synap... . Here’s a link to the GitHub site: https://github.com/kunal333/E2ESynapseDemo .
Based upon feedback from users and our own to-do list, we have released new updates to the solution:
- New 2018 Data is now added to the solution. The solution now has about 148 million rows of real CMS Medicare Part D Healthcare data.
- Design of the Dimensions in Synapse was simplified to improve query performance. The column [Year] in each dimension used to represent every time a granular value existed, and now it represents to most recent year that it existed. See the Azure Update video below for a detailed explanation.
- Logic in Azure Data Factory was consolidated to simplify the data flows and improve performance.
- ‘Smart Narrative’ Power BI AI visual was added to the Summary page of the Power BI report (see video below).
- ‘Q & A’ Power BI AI visual was added to a new page called “Q & A” of the Power BI report (see video below).
- ‘Decomposition Tree’ Power BI AI visual was added to a new page called “Decomp Tree” of the Power BI report (see video below).
- ‘Key Influencers’ Power BI AI visual was added to a new page called “Key Influencers” of the Power BI report (see video below).
- Small Multiples were also added on a new page (see video below).
The video below summarizes some of the changes in Azure:
The video below summarizes some of the changes in Power BI:
So, what’s next? Feedback and recommendations are appreciated and can be provided in the following form. A few potential ideas include expanding to include CMS Medicare Part D reference tables (such as Opioid and Antibiotic lists), integrating new CMS data sets, expanding to new sources of Healthcare Open Data, etc. Please let us know what you’d like to see, and stay tuned!:
