Data Analysis and Related Applications, Volume 2

5A Cluster Analysis Approach for Identifying Precarious Workers

A number of efforts have been made in the past to capture the multifaced phenomenon of employment precarity. However, only a very short body of literature provides quantitative methods that measure the degree and forms of precarity and a way to perform cross-national comparisons. In this chapter, we apply an unsupervised clustering algorithm in order to capture the degree of precarity employment with raw data drawn from the European Labour Force Survey (EU-LFS). The approach engages indicators such as unemployment, part-time employment, temporary contracts, existence of social and health security, among others and is implemented for the case of Greece, with the latest data available at the time, i.e. for the year 2018. However, the methodology could be adopted for other EU member states with no amendments, due to the use of a common questionnaire in the survey for all participating countries. The proposed approach identifies three clusters of workers that exhibit different degrees of precarity. The socio-demographic characteristics of each group are discovered via the cluster centers.

5.1. Introduction

This chapter provides a comprehensive methodology of measuring and defining levels of precarity using raw data drawn from the European Labour Force Survey (EU-LFS). The methodology is implemented using data for the case of Greece, and three levels of precarious workers are identified based on several characteristics ...