Data Driven Identification of Factors Affecting the Job Satisfaction of Programmers

Abstract

In this paper, the factors affecting the programmers' job satisfaction are investigated. We use the recent questionnaire by the StackOverflow team containing 153 questions with more than 50,000 samples from 213 different countries. After some preprocessing, a dataset containing 40,377 samples with 77 factors is prepared. The importance of different factors on job satisfaction are investigated using five different methods including random forest, linear regression, light GBM, correlation analysis, and a hand-designed variance based technique. Finally, different ordered lists obtained by these methods are aggregated using the Borda count algorithm. The results show that factors like career satisfaction, job seeking status, developer type, hours per week to find a new job, and salary are among the most influential factors on programmers' job satisfaction.