Data Smells in Public Datasets

TL;DR

This paper introduces a catalog of data smells, analogous to code smells, to identify early signs of data quality issues in public datasets, aiding data scientists in improving data quality for AI applications.

Contribution

It presents the first systematic catalog of data smells and analyzes 25 datasets to assess the prevalence of these quality issues.

Findings

Identified 14 common data smells in public datasets

Analyzed 25 datasets to evaluate data quality issues

Provided a tool for early detection of data problems

Abstract

The adoption of Artificial Intelligence (AI) in high-stakes domains such as healthcare, wildlife preservation, autonomous driving and criminal justice system calls for a data-centric approach to AI. Data scientists spend the majority of their time studying and wrangling the data, yet tools to aid them with data analysis are lacking. This study identifies the recurrent data quality issues in public datasets. Analogous to code smells, we introduce a novel catalogue of data smells that can be used to indicate early signs of problems or technical debt in machine learning systems. To understand the prevalence of data quality issues in datasets, we analyse 25 public datasets and identify 14 data smells.