What Are The Technical Skills that Affect the Annual earnings in the US IT Industry, and how Machine Learning Algorithms is Ranking The Important Skills

Chenke Bai

Undergraduate Honors Thesis

What Are The Technical Skills that Affect the Annual earnings in the US IT Industry, and how Machine Learning Algorithms is Ranking The Important Skills Public Deposited

Analytics

Download PDF

Citations

Citeable URL: https://scholar.colorado.edu/concern/undergraduate_honors_theses/9593tw45v

Abstract

This study uses multiple different machine learning models to investigate how different skills are affecting the earnings in the IT field. To investigate that we decided to take 7 different machine learning approaches to help identify the factors that affect IT workers' earnings by applying feature importance analysis to the best performing machine learning model, based on the evidence we saw from exploring similar literature on the internet. The topic of study that researches skill-based salary analytics and prediction in the most important skill sets that affect people’s salary in the US IT field is under-researched. The limited study in IT skill and respective salary rate provides us a great opportunity to apply training on selected machine learning models that can be then used to predict IT workers’ salaries based on their skill sets. The data used in this study is StackOverflow 2021 Developer Survey data. This company conducts these surveys annually to gather users’ information so StackOverflow can improve its service. Our choice of computational tool is Python 4.0, the machine learning model was built using Scikit learn library. By using Scikit learn we can adopt 7 machine learning(ML) algorithms and UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction technique into this study. After in-depth analysis, we got the highest testing accuracy of 80%, and around 50% of the salaries were predicted. At the end of this study, we discovered that even though technical IT skills are very important in affecting peoples’ earnings, years of professional technical experience always outweigh a specific technical skill. Furthermore, if we wish to look at skills to see how they affect the salary, we need to look at multiple skills together as a whole. The findings of this study can be used for developing a skill-based salary calculator that allows future IT workers to project their potential salary. By entering different skill combinations and their experience they can see which skills are more likely to yield a higher salary.

Creator