Coderbot: um assistente virtual para apoiar a aprendizagem de programação por meio de worked examples

Abstract

The teaching-learning process in programming courses is so complex that the approval rates in these courses are low. This is a challenge frequently reported by educators teaching programming courses. The content taught is often difficult for students to comprehend, as it requires them to develop skills such as abstraction, problem-solving, reasoning, and logical thinking. To foster more effective learning, there has been an increase in the adoption of Active Methodologies by educators to improve student engagement and motivation. Among these methodologies, Example-Based Learning (EBL) is an effective and efficient approach to teaching students new problem-solving skills. EBL is supported by the Worked Examples (WE) technique, which presents a problem, the steps to solve it, and, finally, a complete solution. In addition, educational technologies have been evolving daily, serving as an excellent means of attracting and engaging students. One of these emerging technologies is chatbots, which are applications that simulate conversati- ons through natural language and/or predefined data. In this context, this master’s thesis presents a virtual assistant called CoderBot, designed to support programming learning in undergraduate courses. To achieve this goal, CoderBot applies EBL with the support of the WE technique, using both correct and incorrect examples. An experimental study was conducted to assess the feasibility and performance of CoderBot as a virtual assistant. Based on the experiment results, CoderBot proved to be an efficient and practical tool for supporting programming learning, promoting comprehension, task resolution, and a comfortable user environment. With good usability and a positive impact on students’ motivation and performance, the tool facilitates learning. However, there is room for improvement in aspects such as innovation and support for beginners.