Download This PaperOn AI Perspectives in Surveys Data Analysis: A Personal Overview
11 Pages Posted: 3 Apr 2025 Last revised: 30 Mar 2025
Date Written: February 05, 2025
Abstract
Surveys are based on structured questionnaires that generate data that can be statistically analysed. The analysis can be performed with a range of models such as multinomial regression, log linear models, structural equations and Poisson regressions or graphical models such as decision trees and Bayesian networks. Artificial intelligence (AI) and machine learning (ML) are gaining presence in all aspects of research and data analysis. Strictly speaking, ML is a subfield of AI about the algorithms and statistical tools that allow computer systems to perform specific tasks without explicit instructions. One side effect of this evolution is an expanded interest in data driven studies including topics such as data integration, alternative scenario analysis and predictive analytics. This paper is a personal retrospective view of experience gained by applying statistics and analytics to surveys, with an emphasis on the past few years. It aims to show how AI and ML merge with statistics in a modern analytic environment that provides new perspectives on the analysis of survey data. The novelty in this overview is an emphasis on survey data analysis leveraging predictive analytics. The paper consists of seven sections designed to provide an overview, without technical details. It covers applications to surveys of the information quality framework, Bayesian networks, decision trees and text analytics. This list is meant to present a broad but not comprehensive coverage of AI applications to surveys. For details on implementation examples, the reader is directed at references.
Keywords: Artificial intelligence, machine learning, survey design, survey analysis, enumerative studies, analytic studies
Suggested Citation: Suggested Citation