Aleksander Fabijan

#Data #ABtesting

[Aleksander Fabijan @afabijan]

Preprint of our @ICSEconf #icse17 paper: https://t.co/9g41F0D0U9 The Evolution of Continuous Experimentation #data #datadriven #casestudy

[Aleksander Fabijan @afabijan]

We started ExP blog in June 2020 and published twelve articles in year one (see below). What A/B testing topics are you interested in learning more about (and want to see in year two)?
#abtesting

https://t.co/vRCV5mbN6t

[Aleksander Fabijan @afabijan]

"Ensuring we don’t harm reliability, performance, and user experience is the most important function of A/B testing during this time." Do you agree or have any other learnings that you observed in the area of A/B testing recently? https://t.co/rO3ZA1bGmb

[Bill Gates @BillGates]

This impressive new tool from @IHME_UW will help hospitals, policymakers, and the general public better understand and prepare for the #COVID19 response in the U.S. https://t.co/tTd5tEtTRc

[Aleksander Fabijan @afabijan]

@nytimes Idea to improve the trustworthiness of news articles using simple ML: USE NLP and alert the composer whenever the context about one country includes a capital city of another country.

[Aleksander Fabijan @afabijan]

Data Quality: Searching for a trustworthy newspaper to update me on the nuclear disarmament talks happening in Sweden today. Still searching. @nytimes the capital of #Sweden is Stockholm. #Helsinki is in Finland.

[Ronny Kohavi @ronnyk]

KDD 2019 paper: Diagnosing Sample Ratio Mismatch in Online Controlled Experiments at https://t.co/E3h6oWBtQq .

If you experiment, you MUST check for SRMs, or your system is not trustworthy. Period.

Trivial implementation I shared: https://t.co/NI2ULsaCDB

[Aleksander Fabijan @afabijan]

@xLeitix Haha, I'm not using twitter too often :). Enjoy the paper. The Safe Velocity paper is also up on https://t.co/YdyNA1KQx2 now!

[Aleksander Fabijan @afabijan]

@xLeitix https://t.co/bYqaezbwWn :)

[Aleksander Fabijan @afabijan]

Experimentation Epidemic in Software Industry – Are You Infected? #ABTesting #HiPPO https://t.co/7SMUEtyPBL

[Aleksander Fabijan @afabijan]

How good are companies at growing their experimentation programs and running #abtests at scale? Check our latest survey paper presented earlier today at SEAA'18 to answer this question! https://t.co/jfgQJaBV9c #experimentation #OEC #HiPPO #ABtesting

[Aleksander Fabijan @afabijan]

@p_diebold Good luck!

[Aleksander Fabijan @afabijan]

A great example of why you should run controlled experiments and not trust observational studies. #abtesting https://t.co/REdNt2RWFR

[Aleksander Fabijan @afabijan]

The only type of #HiPPO that we can trust. For everything else let's use #data.

[Aleksander Fabijan @afabijan]

@ronnyk @lukasvermeer

[Aleksander Fabijan @afabijan]

I specified on https://t.co/rvQohEBLMe search that we will be traveling with a few months old baby, and yet one of the first results is a bed in a hostel room with 8 bunk beds.Designing a good #OEC is hard. Can the other guests handle my baby cry? #ABtesting

[@PROFES_conf @PROFES_conf]

Let the #profes17 skiing day start!

[Aleksander Fabijan @afabijan]

Watching The Bing Bang Theory years ago finally paid off! Singing #softkitty shortens the time needed for my 3months old Annabelle to fall asleep. Results are statsig. Ongoing experiment with walking her around the room\not (control) pending @lukasvermeer

[Aleksander Fabijan @afabijan]

Captain of today’s @KLM fl. #kl1124 greeted the passengers with flight details before(!) the boarding started. #bonusearned #CustomerService

[Aleksander Fabijan @afabijan]

@matseinarsen 6. Evaluation yes, but most importantly, setting the direction in the first place.