Application of Data Analytics Techniques for Decision Making in the Retrospective Stage of the Agile Scrum Methodology

The Scrum retrospective is the last activity within the Scrum methodology, and it is the previous one to Scrum planning within the activities of the method. To propose improvements for future Sprints, the team performs frequent inspection and adaptation processes analyzing what has worked and what has not worked with respect to the current Sprint and the different related tasks such as the estimation of the Product Backlog, assignment of Story Points and prediction of team velocity. These practices are usually supported by empirical team approaches, implying taking risks that can cost development resources. The adoption of data analytics within the business world has gained relevance in recent years, since through this it is possible to process data for decision-making based on predictive models. Today these models are used based on the management and processing of historical data, using data analytics and artificial intelligence techniques. Taking advantage of the fact that software project managers publicly offer Datasets with information on their Sprints and their characteristics, these could be used with data analytics, in conjunction with supervised learning models, both prediction and classification to support a more efficient decision making, thereby achieving an improvement in the process with a more efficient use of resources.