NEXUS: On Explaining Confounding Bias

Brit Youngmann, Michael Cafarella, Yuval Moskovitch, Babak Salimi

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


When analyzing large datasets, analysts are often interested in the explanations for unexpected results produced by their queries. In this work, we focus on aggregate SQL queries that expose correlations in the data. A major challenge that hinders the interpretation of such queries is confounding bias, which can lead to an unexpected association between variables. For example, a SQL query computes the average Covid-19 death rate in each country, may expose a puzzling correlation between the country and the death rate. In this work, we demonstrate NEXUS, a system that generates explanations in terms of a set of potential confounding variables that explain the unexpected correlation observed in a query. NEXUS mines candidate confounding variables from external sources since, in many real-life scenarios, the explanations are not solely contained in the input data. For instance, NEXUS might extract data about factors explaining the association between countries and the Covid-19 death rate, such as information about countries' economies and health outcomes. We will demonstrate the utility of NEXUS for investigating unexpected query results by interacting with the SIGMOD'23 participants, who will act as data analysts.

Original languageEnglish
Title of host publicationSIGMOD 2023 - Companion of the 2023 ACM/SIGMOD International Conference on Management of Data
Number of pages4
ISBN (Electronic)9781450395076
StatePublished - 4 Jun 2023
Event2023 ACM/SIGMOD International Conference on Management of Data, SIGMOD 2023 - Seattle, United States
Duration: 18 Jun 202323 Jun 2023

Publication series

NameProceedings of the ACM SIGMOD International Conference on Management of Data


Conference2023 ACM/SIGMOD International Conference on Management of Data, SIGMOD 2023
Country/TerritoryUnited States


  • aggregated SQL queries
  • confounding bias
  • knowledge graphs

All Science Journal Classification (ASJC) codes

  • Software
  • Information Systems


Dive into the research topics of 'NEXUS: On Explaining Confounding Bias'. Together they form a unique fingerprint.

Cite this