Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
@fugu13
russell.duhon@saltbox.com
Russell Duhon, CTO of Saltbox, Makers of WaxLRS
I’ve been tweeting with
#xapi, join me! I’d love to
read your comments later.
Running an Enterprise
Implementation of xAPI,
explained succinctly, I hope.
The past is enterprise…
…but so’s the future.
Enterprise is…
• business relevant,
• in the flow of work,
• where the traces of learning manifest,
• and a reporting driv...
Diversity of business systems
+ flexibility of xAPI
=
data modeling matters a lot
Very Important Modeling Considerations
• Identity, identity, identity
• Start with core events
• Verb+Activity
• Context a...
Some Data Modeling Steps
• Model Data
• Make Example Data
• Cross Check Data Model with Developer/IT
• Choose Test Dataset...
Modeling Data Depends
on How You Move It
Moving Data, Most to Least Preferred
• Extract, Transform, and Load
• full access APIs via Zapier or similar
• Plugin
• fu...
Very Important Technical Tidbits
• Don’t run stuff in web browsers
• Queue (SQS, perhaps)
• Idempotency by setting Timesta...
Some Technical Steps
• Introduce errors at each stage
• Test on all available historical data
• Deploy!
• Define & test au...
@fugu13 — russell.duhon@saltbox.com
Russell Duhon, CTO of Saltbox, Makers of WaxLRS
Upcoming SlideShare
Loading in …5
×

Enterprise Implementation talk at xAPI Camp - Amazon

2,235 views

Published on

How to run enterprise system integrations for the Experience API, presented at the xAPI Camp held at Amazon on July 21st.

Published in: Data & Analytics
  • Be the first to comment

Enterprise Implementation talk at xAPI Camp - Amazon

  1. 1. @fugu13 russell.duhon@saltbox.com Russell Duhon, CTO of Saltbox, Makers of WaxLRS
  2. 2. I’ve been tweeting with #xapi, join me! I’d love to read your comments later.
  3. 3. Running an Enterprise Implementation of xAPI, explained succinctly, I hope.
  4. 4. The past is enterprise…
  5. 5. …but so’s the future.
  6. 6. Enterprise is… • business relevant, • in the flow of work, • where the traces of learning manifest, • and a reporting driver.
  7. 7. Diversity of business systems + flexibility of xAPI = data modeling matters a lot
  8. 8. Very Important Modeling Considerations • Identity, identity, identity • Start with core events • Verb+Activity • Context aids richness • Specificity saves you later
  9. 9. Some Data Modeling Steps • Model Data • Make Example Data • Cross Check Data Model with Developer/IT • Choose Test Dataset • Transform & Explore Test Dataset
  10. 10. Modeling Data Depends on How You Move It
  11. 11. Moving Data, Most to Least Preferred • Extract, Transform, and Load • full access APIs via Zapier or similar • Plugin • full access API connector • synchronization portal for individually authorized APIs
  12. 12. Very Important Technical Tidbits • Don’t run stuff in web browsers • Queue (SQS, perhaps) • Idempotency by setting Timestamp & Statement ID • Don’t discard errors • Use batch & concurrent operations, but set limits • Edge cases matter
  13. 13. Some Technical Steps • Introduce errors at each stage • Test on all available historical data • Deploy! • Define & test automated monitoring procedures
  14. 14. @fugu13 — russell.duhon@saltbox.com Russell Duhon, CTO of Saltbox, Makers of WaxLRS

×