Welcome to the last study guide in my Einstein Certification series! For this section, we will briefly cover topics on migration, XMD, and dashboard performance. For notes and resources on enabling Einstein Analytics please see the Security guide — it was easier for me to combine that information there.
- Trailhead Badges
-Any from this Trailmix you haven’t completed yet!
- Documentation in Exam Guide
-Deploy your Changes
-Analytics Migration, Packaging, and Distribution
-Extended Metadata (XMD) Overview
- Einstein Analytics Training Videos
-Videos 22, 24, 32
- External Resources
-Colors, labels, values — oh my!
-Einstein Analytics: Getting started with XMD
-8 Tips to Increase Dashboard Performance
-Einstein Analytics: deployment with Change Set
Migrating EA Assets
Many ways to migrate Einstein assets from Sandbox to production or external orgs.
- Manually copy and paste dashboard JSON from sandbox to production; download then upload dataflow JSON and XMD files
- Change Sets — You can migrate Analytics Assets the same way you would migrate things like fields or workflows in Salesforce: through the Outbound Change Set tool.
— All the components start with Analytics: Analytics Application/Analytics Dashboard/Analytics Dataset/Analytics Recipe, etc.
— Must have Analytics Admin permissions to create the package, but not required to deploy it.
— Change sets will not move data; must re-run any dataflows you move over.
- ANT Script
— Conditional formatting will not migrate through ANT unless you manually add the XMD asset to the deployment
- Managed Packages
— Can use managed packages of apps/dashboards/lenses etc. to migrate to other users or organizations outside your company
— *Unmanaged packages are not currently GA for Einstein Analytics
- AppExchange Packages
— Can also publish EA managed packages through the AppExchange; will go through a security review first
- Metadata API
Extended Metadata — set of instructions on how the data is visualized; structure in JSON format.
- XMD options within each dataset:
— Change the display labels for fields
— Hide fields from explorer and dashboard designer
— Change how the values appear or what colors default to those items
— Customize how a measure appears: whether it defaults to a number, currency, percent or define your own custom format
— Multiply measures by a factor. Ex. *100 to convert a decimal to a percent
— Specify what fields show up by default when a user explores a lens using the dataset
— Add commonly used derived measures and dimensions to your dataset
— Enable actions on fields within the dataset
— Define the first day of the week for a calendar year
— Change the label for the row count measure
- When a step utilizes multiple datasets, the results are formatted using the XMD of the first loaded dataset
Dashboard Performance Tips
- Dashboard inspector — Tool inside dashboards that checks the performance of the dashboard as well as individual steps
— Tells you which steps are running long and not optimized
— *Also helpful for troubleshooting when a widget isn’t showing expected results! Use “original query” versus “final query” to investigate
- Tips to improve performance
— Remove unused steps
— Minimize bindings; use Connect Data Sources instead
— Reduce number of steps and consider moving some into a different dashboard.
— Use global filters and reduce overall number of filters
— Utilize the minimum decimals required
— Avoid filtering using measures (they aren’t indexed)
— Only use SAQL when you absolutely have to; perform complex calculations inside the dataflow instead of SAQL
— Use dashboard pages to break up content
— Reduce the size of your dataset and detailed data tables
— Make sure to test with large datasets
- Things you can’t control:
— Browser speed
— Internet speed of the end user
This concludes my study notes — best of luck on your test!