This study develops and validates an innovation management framework based on the integration of dynamic capabilities and multidimensional innovation process factors. A mixed-methods approach was employed, combining a quantitative survey of 267 respondents from the Porto Digital ecosystem with qualitative interviews conducted in 10 companies. Confirmatory factor analysis (CFA) was used to validate the proposed measurement model, confirming the reliability and validity of the identified constructs. The qualitative findings reinforced the contextual relevance of the framework and provided insights into managerial perceptions of innovation capabilities. The validated framework consolidates ten organizational factors into four hierarchical layers, offering a structured approach to assessing and strengthening innovation management capacity. This study contributes to the literature by proposing an empirically validated model that addresses gaps related to dynamic capabilities and integrated innovation processes. Practical implications are also discussed, providing managers with a diagnostic tool to support strategic innovation initiatives.

The delivery sector, already on an upward trend prior to the Covid-19 pandemic, experienced a significant boost in demand for its services during the pandemic. However, the most expensive aspect for companies in this domain is the 'last mile' delivery costs, which can account for up to 45% of total logistics expenses. In light of this, our paper presents a study aiming to leverage this market trend to reduce company costs. We propose a hybrid model combining trucks and drones for last-mile delivery and conduct a market analysis for a company interested in developing this model. Drone deliveries are anticipated to grow at an annual rate of approximately 45%, outpacing the overall sector. This market outlook presents a compelling opportunity for entrepreneurs: the market is nascent, with a noticeable upward trend in both demand and investment. The truck-drone hybrid model, being up to 70% faster and 60% less expensive compared to traditional methods, offers dual benefits: enhanced customer satisfaction with lower expenditure, and reduced operational costs for companies while improving service quality. Despite pending regulatory challenges, the prevailing trend in technology advancements and major retailers' interest suggests a favorable outlook for legal approvals. As a novel technology, it naturally encounters some skepticism. Furthermore, drone usage has positive environmental implications, whether through alternative energy sources or traffic reduction, aligning well with the current ESG-driven market trends.

Electrodermal Activity (EDA) has emerged as a valuable physiological measure in educational research, providing insights into emotional and cognitive engagement. This study investigates the variation in students' EDA responses under traditional lecture-based and active learning conditions. Data were collected from eight university students using electrodermal resistance sensors during instruction on the Single Minute Exchange of Die (SMED) methodology. Each participant engaged in both instructional modalities within a controlled environment. The EDA signals were analyzed using statistical techniques including paired t-tests and Mann-Whitney tests. Results indicated that most students exhibited significantly higher EDA levels during active learning sessions, suggesting increased arousal and engagement. However, individual differences, including potential non-responsiveness and gender-based variability, were also observed. These findings underscore the potential of EDA as a realtime, non-invasive tool for assessing instructional effectiveness and student engagement, offering implications for the design of adaptive and student-centered learning environments.

A crescente globalização e competitividade têm impulsionado empresas a buscar otimização de processos através do uso de tecnologias avançadas, como o aprendizado de máquina (machine learning). Este estudo investiga a aplicação de modelos de aprendizado de máquina para a predição da temperatura do rotor de motores síncronos de ímã permanente (PMSM), utilizando uma base de dados real coletada pela Universidade de Paderborn. Foram avaliados cinco modelos: regressão linear, regressão polinomial (ordem 2 e 3), árvore de decisão e aprendizado profundo. Os resultados indicaram que a regressão polinomial de terceira ordem apresentou a melhor performance, com uma variância explicada de 93%, MSE de 22.54 e RMSE de 4.74. A capacidade de prever a temperatura do rotor com precisão pode melhorar significativamente a manutenção preditiva, a eficiência operacional e a segurança dos motores PMSM. O estudo também destaca a importância do tratamento de dados e ajuste de hiperparâmetros para a melhoria da performance dos modelos.

A sociedade contemporânea é responsável por gerar várias fontes de dados, que são convertidas em grandes volumes de Big Data. Esses dados são armazenados em Data Centers e, mais recentemente, em Data Lakes, que possuem uma capacidade ainda maior de armazenamento e que são capazes de receber grandes volumes de dados variados. O cotidiano do processamento de dados é feito por intermédio de softwares como Hadoop, Hive, SAS e SQLMS, juntamente com outros, e a integração dessas diversas fontes é complexa no ambiente empresarial. A automação, especialmente pela linguagem Python, pode reduzir variáveis no processo da eficiência operacional, uma vez que é viável diminuir o tempo de processamento e erradicação do número de erros. O presente trabalho visa a investigar os benefícios advindos da robótica no processo de unificação e automação dos dashboards de uma grande empresa do setor financeiro do Brasil, tendo em vista a eficiência e a confiabilidade das tarefas.