Webinar: DataOps For Beginners – 2024

If you’ve ever heard (or had) these complaints about speed-to-insight or data reliability, you should watch our webinar, DataOps for Beginners, on demand.

Written by Gil Benghiat on October 30, 2024

DataOps Data Observability DataOps TestGen On-Demand Webinar Open Source

Key points

DataOps came out of two problems the presenters lived through: shipping wrong data to thousands of people for a month before anyone noticed, and being told "I thought this should be two hours, not two weeks."
DataOps is three things at once — Agile for the people and the process, DevOps automation for the technical environment, and lean-manufacturing measurement, including statistical process control, for the operation.
There are two pipelines to automate and test. In the value pipeline (production) the code holds still and the data changes; in the innovation pipeline (development) the data holds still and the code changes. The same test often works in both, which is where the savings come from.
Production tests are not pass/fail. Some data is bad enough to stop the line, some warrants a warning or a release note, and some is just a measurement you keep — but keep the history either way, so you can chart the trend and set a real threshold.
Limit work in progress. The study Gil cited puts the cost of context switching at roughly 20% of your time for each additional project you take on.
The seven steps are: orchestrate the two journeys, add tests and monitoring, use version control, branch and merge, run multiple environments, reuse and containerize, and parameterize.

“That should take two hours, not two months. Can’t your Data & Analytics Team go any faster?”

“The executives’ dashboard broke! The data’s wrong! Can I ever trust our data?”

If you’ve ever heard (or had) these complaints about speed-to-insight or data reliability, you should watch our webinar, DataOps for Beginners, on demand.

DataKitchen’s VP Gil Benghiat breaks down what DataOps is (spoiler: it’s not just DevOps for data) and how DataOps can take your Data & Analytics factory from clunky, junky, and funky to state-of-the-art with quality control and speed.

Specifically, Gil covers: