Electricity expense management presents significant challenges, as this resource is susceptible to various influencing factors. In universities, the demand for this resource is rapidly growing with institutional expansion and has a significant environmental impact. In this study, the machine learning models long short-term memory (LSTM), random forest (RF), support vector regression (SVR), and extreme gradient boosting (XGBoost) were trained with historical consumption data from the Federal Institute of Paraná (IFPR) over the last seven years and climatic variables to forecast electricity consumption 12 months ahead. Datasets from two campuses were adopted. To improve model performance, feature selection was performed using Shapley additive explanations (SHAP), and hyperparameter optimization was carried out using genetic algorithm (GA) and particle swarm optimization (PSO). The results indicate that the proposed cooperative ensemble learning approach named Weaker Separator Booster (WSB) exhibited the best performance for datasets. Specifically, it achieved an sMAPE of 13.90% and MAE of 1990.87 kWh for the IFPR–Palmas Campus and an sMAPE of 18.72% and MAE of 465.02 kWh for the Coronel Vivida Campus. The SHAP analysis revealed distinct feature importance patterns across the two IFPR campuses. A commonality that emerged was the strong influence of lagged time-series values and a minimal influence of climatic variables.

Educational institutions are essential for economic and social development. Budget cuts in Brazil in recent years have made it difficult to carry out their activities and projects. In the case of expenses with water and electricity, unexpected situations can occur, such as leaks and equipment failures, which make their management challenging. This study proposes a comparison between two machine learning models, Random Forest (RF) and Support Vector Regression (SVR), for water and electricity consumption forecasting at the Federal Institute of Paraná-Campus Palmas, with a 12-month forecasting horizon, as well as evaluating the influence of the application of climatic variables as exogenous features. The data were collected over the past five years, combining details pertaining to invoices with exogenous and endogenous variables. The two models had their hyperpa-rameters optimized using the Genetic Algorithm (GA) to select the individuals with the best fitness to perform the forecasting with and without climatic variables. The absolute percentage errors and root mean squared error were used as performance measures to evaluate the forecasting accuracy. The results suggest that in forecasting water and electricity consumption over a 12-step horizon, the Random Forest model exhibited the most superior performance. The integration of climatic variables often led to diminished forecasting accuracy, resulting in higher errors. Both models still had certain difficulties in predicting water consumption, indicating that new studies with different models or variables are welcome.

The automation of urban layout generation has evolved significantly, yet existing evolutionary and deep learning methodologies often struggle to balance morphological diversity with strict topological constraints, such as guaranteed street access and adaptability to irregular terrains. This work presents a novel unsupervised framework for seamlessly adaptable and adjustable urban subdivision design, introducing two constructive algorithms: Parcel k-Means and Parcel Binary Space Partitioning (BSP). Unlike "black-box" Generative Adversarial Networks (GANs) or computationally expensive Genetic Algorithms, our proposal offers a "Glass Box" approach that ensures 100% constructability, reproduction and geometric validity by design. The method natively handles non-convex, inclined, and complex topographies without requiring post-processing or repair mechanisms. Experimental results demonstrate speed improvements, reducing generation times to seconds (less than 1 second) while fully eliminating common errors like landlocked parcels. By utilizing simpler input data and providing interpretable parameter control, this framework allows planners to rapidly explore diverse, engineering-compliant layout alternatives, marking the first application of k-Means clustering to urban subdivision in the literature. The proposed BSP also differs from the existing literature due to the employed heuristics and by being capable of working with irregular quadrilaterals as input land. Both approaches can also generate different results by varying just a single parameter such as the seed variable, which we call multi-option.

Com o objetivo de destacar a importância das técnicas de visualização de dados, como storytelling de dados e dashboards, no processo de tomada de decisão na gestão pública, o presente trabalho busca descrever como essas técnicas foram implementadas no software Prefeitura Interativa. O ambiente de apoio à decisão foi elaborado para auxiliar a administração pública municipal na gestão de políticas públicas e na busca por soluções mais eficientes para as demandas que afetam a população local. Por meio de uma abordagem colaborativa, o software permite a aplicação de técnicas específicas de coleta e visualização de dados para melhorar a análise e interpretação de informações relevantes para a tomada de decisão.