Get Help From Experts in Data Science and Mathematics
ePortfolios |STEM Gender Equality | ZOOM | Slack | Spreading Mathematical Happiness

You can support MathsGee with your DONATION

1 like 0 dislike
29 views
What variants of data pre-processing do the Uber experimentation team apply  to improve the robustness and effectiveness of their A/B analyses?
by Wooden (1,460 points) | 29 views

1 Answer

1 like 0 dislike
Best answer

Three variants of data preprocessing are applied to improve the robustness and effectiveness of Uber's A/B analyses:

  • Outlier detection removes irregularities in data and improves the robustness of analytic results. They use a clustering-based algorithm to perform outlier detection and removal.
  • Variance reduction helps increase the statistical power of hypothesis testing, which is especially helpful when the experiment has a small user base or when they need to end the experiment prematurely without sacrificing scientific rigor. The CUPED Method  leverages extra information they have and reduces the variance in decision metrics.
  • Pre-experiment bias is a big challenge at Uber because of their diversity of users. Sometimes, constructing robust counterfactual via mere randomization just doesn’t cut it. Difference in differences (diff-in-diff) is a well-accepted method in quantitative research and we use it to correct pre-experiment bias between groups so as to produce reliable treatment effects estimation.
by Wooden (1,460 points)

Related questions

0 like 0 dislike
1 answer
1 like 0 dislike
1 answer
1 like 0 dislike
1 answer
0 like 0 dislike
1 answer
1 like 0 dislike
1 answer

Welcome to MathsGee Q&A Bank, Africa’s largest personalized FREE Study Help network that helps people find answers to problems and connect with experts for improved outcomes.


Get help from experts - simply ask your question


You can Support MathsGee with your DONATION

Enter your email address:

13,102 questions
10,327 answers
101 comments
11,180 users