Impact of COVID-19 on Productivity of Software Engineers: A Comparative Analysis of work from home (WFH) and work from office (WFO) environment using Machine Learning

The abstract of this research study encapsulates an exploration of the transformative impact of the COVID-19 pandemic on the software and IT industry, specifically concerning the work practices of software engineers. With a significant shift towards remote work arrangements, both partial and full-time, this study examines the resultant changes in productivity. Through a comparative analysis, the research contrasts the preferences and outcomes of Working from Home (WFH) and Working from Office (WFO) modes among software engineers. Data was gathered through a questionnaire distributed to 174 professionals in the Punjab region, and subsequent analysis employed diverse data processing techniques, including machine learning algorithms. The study’s primary goal is to identify the preferred work setup for software engineers, assessing the accuracy of various classifiers in predicting these preferences based on collected data. Notably, the Decision Tree and Random Forest classifiers showcased impressive accuracies of 96.23 %, reinforcing the models’ robustness. The Support Vector Machine achieved an accuracy of 90.56 %, while the Logistic Regression model demonstrated an accuracy of 88.67 %, offering valuable insights. The Multinomial Naive Bayes algorithm yielded an accuracy of 77.36 %, contributing to understanding work preference dynamics. Lastly, the K Neighbors model demonstrated an accuracy of 67.92 %, suggesting potential limitations in capturing nuanced patterns. The abstract briefly overviews the research’s objectives, methodology, and critical findings, including the accuracy rates of various classifiers.