Code Is My Copilot (at #BITSS2019) (@Code_Copilot) header
Code Is My Copilot (at #BITSS2019) (@Code_Copilot) profile picture

Reducing human error in data analysis by helping researchers find others to independently audit research code. See posting guidelines (https://teapawt.org/t.co/hyEE2n69yC).

Stanford, CA
Joined November 2019
Photo by Code Is My Copilot (at #BITSS2019) (@Code_Copilot)
Code Is My Copilot (at #BITSS2019) retweeted
Seeking @Code_Copilot for JudgementAndDecisionMaking experiment wt @giladfeldman, prefer by 25Feb2020. 150lines data prep + 1450 lines analysis (large chunks of it are empty rows). Background needed: R, Rmarkdown. Background useful: prop. tests, rudimentary BayesianAnlysis
4
2
We'll be at #BITSS2019! Come say hi to @_MMathur and @Tom_Hardwicke and grab a free decal.
2
A wonderful first example of copiloting by @jpheffne, @giladfeldman, & @PRASAC!
Replying to @Code_Copilot
@Code_Copilot works, and is terrific! Went over what @jpheffne sent @PRASAC after reviewing our manuscript's Rmarkdown, & it's wonderful. Truly. That's what science should be. Open, reproducible, collaborative, positive. Contribute to your community! Thanks @jpheffne!
2
7
And our first intrepid copiloting offer! Thanks, Joey!
Cool concept @Code_Copilot! I’d be happy to help review R code
1
2
Our second copiloting request! Who would like to help with this?
Seeking @Code_Copilot for JudgementAndDecisionMaking experiment wt @giladfeldman , prefer by 10Dec2019. 150lines data prep + 920 lines analysis(a large portion of it is empty rows). Background needed: R, Rmarkdown. Background useful: ind T-tests, rudimentary BayesianAnlysis
3
4
Replying to @sankar_alagapan
We hear you on the deep fear of analysis errors. @_MMathur has had literal nightmares to this effect...
Replying to @phillip4real
Per our posting guidelines, we would like this information included in initial requests.
Our inaugural copiloting request! Who can help copilot this simulation study? Someone from #epitwitter?
“Seeking @Code_Copilot for maximum likelihood of a mixture cure model with covariates via simulation, wanted by by 20/11/2019
Replying to @BrettButtliere
Requesters post their own requests by tweeting @Code_Copilot, which we then retweet. See our posting guidelines for more on this.
1
And that should be @Tom_Hardwicke. If only we'd had a copilot to audit our tweet...
6.) Questions? Want more details? See our documentation and FAQ (shorturl.at/mnIP9)!
1
1
6
5.) We hope this idea will catch on and survive. This largely depends on your own engagement and sharing. If you support, please consider retweeting, following, posting a request, or offering to copilot.
1
1
9
4.) If you’re interested in seeing these requests and potentially being a copilot, please follow us. No strings or commitments required. To offer your copiloting skills, just respond directly to the request tweet. You and the requester can take it from there.
2
1
6
3.) “Seeking @Code_Copilot for human-robot interaction experiment, wanted by 1/1/2020. 450 lines data prep + 620 lines analysis. Background needed: R, tidyR, linear mixed models. Background useful: mouse-tracking data.”
1
1
8
2.) To find a copilot for your study, send us @Code_Copilot a tweet following our posting guidelines (shorturl.at/mnIP9). We will then retweet to our followers. Potential copilots can respond directly. For example:
1
1
6
1.) @Code_Copilot (by @_MMathur, @TomHardwicke, & @BalazsAczel) is a no-frills, beta platform to help you, a researcher in any discipline, to find another researcher (your “copilot”) to carefully read and re-run your data prep and analysis code to help check for errors.
3
2
3
16
Have you ever found a critical error in your data analysis code? We all have. One of the best ways to catch errors is through independent third-party auditing. This new account is designed to help; more below. (Please retweet.)
3
120
14
178