Merelda Wu
Merelda Wu
๐„๐„™๐„๐„™๐„๐„™๐„๐„™๐„๐„™๐„๐„™๐„๐„™๐„™๐„๐„™๐„™๐„๐„™๐„
Nov 10, 2020 1 min read

Why data quality is key to successful ML Ops

img

Are you testing your data in your data pipeline ๐Ÿงช?

We know that garbage in = garbage out, and we also know that data scientists (unfortunately) spend 65% of their time on data cleaning. Then, surely, data validation is an even more important step to make sure all that effort is not being eroded by poor data quality?

From the post: โ bad data can backfire twice โ€” first when training predictive models and second in the new data used by that model to inform future decisions. โž

๐Ÿ’ฃ So, are you testing your data?

If not, check out this great post by Great Expectations on #datavalidation and #MLOps. Link in comment.


๐ŸŒŸ ย  Original Post by Great Expectation on: Why data quality is key to successful ML Ops