Create confidence_intervals module #25
Conversation
This creates a confidence_intervals module with a function that calculates Wilson score binomial proportion confidence intervals.
|
Hi @tphan-splk, I noticed you're the last person to commit to this repo. Would you be able to review this PR? Thanks! |
|
Hi @jcostello-atlassian and thank you for the suggestion! We will need to spend some time on the organization for this type of functionality and details on the signature. We loosely try to mimic NumPy where possible. We also need a general policy/procedure re: external contributions as this GitHub repository merely reflects what is deployed in Splunk Infrastructure Monitoring (fka SignalFx) realms and is intended as documentation for both the "SignalFlow standard library" as well as the SignalFlow language itself; there is no pipeline/mechanism going from this repository to any deployment. |
|
Sure, happy to refactor to align things more. I was lightly using https://www.statsmodels.org/dev/generated/statsmodels.stats.proportion.proportion_confint.html as inspiration, but didn't realize the NumPy precedent. Is there a specific function signature you can share that we can emulate here? As for the deployment pipeline and external contributions, is that something realistically in scope or feasible at this time? I wasn't aware the repo wasn't set up for external contributions, and just wanted to share some functionality for others that's been helpful for us. Also importing the function from the stats library would allow us to DRY things up a bit on our end 😄. So yeah, happy to touch things up if there's a path to deployment, but otherwise am also fine with leaving this as a feature request for now. |
|
Hey @joe-sfx, just wanted to follow up here. Would it be preferable if I created a feature request for SignalFx/Splunk? |
|
Hi, sorry for the delay. Yes, please create a feature request. Thank you! |
This creates a confidence_intervals module with a function that calculates Wilson score binomial proportion confidence intervals.