Towards Measurable Motivation in Software Development

Abstract

IT experts are feeling the strain of fast-paced, demanding work, and working measures for improved well-being are increasingly being established. One essential element of well-being is motivation. Motivation is critical in order to maintain smoothly working processes and quality in completed tasks. To detect and intercept situations where motivation begins to decrease, we need means to identify changes in motivation. To automatically detect such changes requires ways to measure motivation levels from data at hand. In this study, we present preliminary results towards defining measurable motivational factors. Our approach relies on emotional indicators of motivational change that can be detected automatically from software repositories. Our results show that the emotional content of commit comments and issue texts, extracted using sentiment analysis and emotion detection, is correlated to varying degrees with collaboration and risk metrics. We point to possibilities for future work on more sophisticated measures with more complex data.