Developing communication skills in collaborative contexts is of special interest for educational institutions, since these skills are crucial to forming competent professionals for today’s world. New and accessible technologies open a way to analyze collaborative activities in face-to-face and non-face-to-face situations, where collaboration and student attitudes are difficult to measure using traditional methods. In this context, Multimodal Learning Analytics (MMLA) appear as an alternative to complement the evaluation and feedback of core skills. We present a MMLA platform to support collaboration assessment based on the capture and classification of non-verbal communication interactions. The developed platform integrates hardware and software, including machine learning techniques, to detect spoken interactions and body postures from video and audio recordings. The captured data is presented in a set of visualizations, designed to help teachers to obtain insights about the collaboration of a team. We performed a case study to explore if the visualizations were useful to represent different behavioral indicators of collaboration in different teamwork situations: a collaborative situation and a competitive situation. We discussed the results of the case study in a focus group with three teachers, to get insights in the usefulness of our proposal. The results show that the measurements and visualizations are helpful to understand differences in collaboration, confirming the feasibility the MMLA approach for assessing and providing collaboration insights based on non-verbal communication.

Classroom teaching methodologies are gradually changing from masterclasses to active learning practices, and peer collaboration emerges as an essential skill to be developed. However, there are several challenges in evaluating collaborative activities more objectively, as well as to generate valuable information to teachers and appropriate feedback to students about their learning processes. In this context, multimodal learning analytics facilitate the evaluation of complex skills using data from multiple data sources. In this work, we propose the use of beacons to collect geolocation data from students who carry out collaborative tasks that involve movement and interactions through space. Furthermore, we suggest new ways to analyze, visualize, and interpret the data obtained. As a first practical approach, we carried out an exploratory, collaborative activity with sixteen undergraduate students working in a library, with bookshelves and work tables monitored by beacons. From the analysis of student movement dynamics, three types of well-differentiated student roles were identified: the collectors , those who go out to collect data from the bookshelves, ambassadors , those who communicate with other groups, and the secretaries , those who stay at their work desk to shape the requested essay. We believe these findings are valuable feedback for the enhancement of the learning activity and the first step towards MMLA-driven Teaching Process Improvement method.

One of the significant challenges in scaling agile software development is organizing software development teams to ensure effective communication among members while equipping them with the capabilities to deliver business value independently. A formal approach to address this challenge involves modeling it as an optimization problem: given a professional staff, how can they be organized to optimize the number of communication channels, considering both intra-team and inter-team channels? In this article, we propose applying a set of bio-inspired algorithms to solve this problem. We introduce an enhancement that incorporates ensemble learning into the resolution process to achieve nearly optimal results. Ensemble learning integrates multiple machine-learning strategies with diverse characteristics to boost optimizer performance. Furthermore, the studied metaheuristics offer an excellent opportunity to explore their linear convergence, contingent on the exploration and exploitation phases. The results produce more precise definitions for team sizes, aligning with industry standards. Our approach demonstrates superior performance compared to the traditional versions of these algorithms.

Effective collaboration in agile software development depends on communication, cooperation, and coordination. While cooperation is often ensured through defined roles and expertise, coordination—how participants interact during collaborative tasks—remains difficult to evaluate objectively. Agile methods promote coordination through techniques like planning poker, yet there is limited empirical evidence of their actual impact on team communication. Multimodal Analytics (MmA), which enables the quantitative analysis of verbal, paraverbal, and non-verbal cues, offers a promising approach to address this gap. This study examines the effect of coordination techniques on team communication using MmA in a controlled experiment. A total of 72 undergraduate students formed 18 teams of four, self-organized to reflect real-life group formation. Each team performed two software effort estimation tasks, one without coordination (ad hoc) and another using planning poker. The study, designed as a controlled experiment, addresses three research questions regarding the impact of coordination on speaking and attention patterns. Audiovisual recordings were analyzed to measure two dependent variables: speaking time and attention time. For each, we calculated three metrics: total time, average time per participant, and standard deviation among participants. Speaking time was measured via speech diarization, while attention time was inferred from participants’ facial orientation during verbal interactions. Results indicate that while planning poker does not significantly alter total speaking or attention time, it fosters a more equitable distribution of speaking time. This suggests an improvement in balanced participation, an important feature of effective collaboration. The study demonstrates MmA’s utility in capturing subtle team dynamics and contributes empirical evidence on the role of coordination techniques in enhancing collaborative communication in agile environments.

Para a tomada de decisão sobre qual estratégia é melhor para manejo de plantas daninhas podem ser usadas ferramentas estatísticas, como análise multivariada, principalmente quando se dispõe de quantidade elevada de dados. Assim, objetivou-se avaliar a viabilidade da utilização da análise multivariada para estudos de seleção da flora de plantas daninhas em decorrência da aplicação de diferentes herbicidas. O experimento foi conduzido em campo, em delineamento experimental blocos ao acaso, com quatro repetições, e a área foi dividida em 32 parcelas com 3 x 5 metros de comprimento, com 8 parcelas por bloco. Os tratamentos experimentais foram: 1-glifosato; 2- 2,4D; 3- glifosato + 2,4D; 4- carfentrazone-ethyl; 5- glifosato + carfentrazone-ethyl; 6- haloxyfop-methyl; 7- cletodim; 8- glufosinato de amônio. Foram feitos os levantamentos da comunidade infestante em cada parcela antes da aplicação e aos 30 dias após a aplicação dos tratamentos. Os dados foram submetidos à análise de agrupamento e componentes principais. As espécies de plantas daninhas mais discriminantes foram Alternanthera tenella e Commelina benghalensis, seguidas das gramíneas Panicum maximum, Cenchrus echinatus e Eleusine indica. Os resultados da análise de agrupamento foram bastante similares àqueles obtidos pela análise dos componentes principais. As técnicas de estatística multivariada foram capazes de agrupar tratamentos químicos de acordo com a composição específica independentemente do índice fitossociológico considerado.