I was wondering what metrics people here are using...
# ask-metaflow
w
I was wondering what metrics people here are using to measure developer productivity. Has anyone come up with a good way to measure success? Metaflow's core philosophy is to optimize for the productivity and the development experience of the data scientist, but how do you measure that to show it's working?
f
hard to do without any baseline but in my case we're migrating to metaflow from sagemaker pipelines so there's a pretty clear story to tell but higher level than dora-style metrics
haven't really been looking at anything but spend but imagine you'd be able to get solid proxies for efficiency through comparing metaflow w the baseline for things like number of runs completed, time to first deploy + number of touchpoints w external teams
guess those are kind of dora-y, but there's a lot to choose from in my case - especially wrt aws bills since sagemaker makes it so easy to overspend while our metaflow environments are running on spot
Without the comparison still could be useful to track what's going on in the metadata store around adoption / velocity - would be nice if there was a dbt package for that
m
Very interesting question! 🙂 Don't have any pearls of wisdom, but would be very interested in others perspectives on this point.
h
Lines of code of course 🤣 (just kidding)
m
more is better? 😉
🤣 1