A Data Analytics/Big Data Framework for Advanced Metering Infrastructure Data.

The Advanced Metering Infrastructure (AMI) data represent a source of information in real time not only about electricity consumption but also as an indicator of other social, demographic, and economic dynamics within a city. This paper presents a Data Analytics/Big Data framework applied to AMI data as a tool to leverage the potential of this data within the applications in a Smart City. The framework includes three fundamental aspects. First, the architectural view places AMI within the Smart Grids Architecture Model-SGAM. Second, the methodological view describes the transformation of raw data into knowledge represented by the DIKW hierarchy and the NIST Big Data interoperability model. Finally, a binding element between the two views is represented by human expertise and skills to obtain a deeper understanding of the results and transform knowledge into wisdom. Our new view faces the challenges arriving in energy markets by adding a binding element that gives support for optimal and efficient decision-making. To show how our framework works, we developed a case study. The case implements each component of the framework for a load forecasting application in a Colombian Retail Electricity Provider (REP). The MAPE for some of the REP's markets was less than 5%. In addition, the case shows the effect of the binding element as it raises new development alternatives and becomes a feedback mechanism for more assertive decision making.

Identificación de la delimitación administrativa de la malaria usando redes neuronales artificiales / Boundary Delimitiation of Malaria using Artificial Neural Networks / Identificação da Delimitação Administrativa da Malária usando Redes Neurais Artificiais

Resumen La metodología de clustering fue utilizada para agrupar tres barrios en Quibdó teniendo en cuenta factores que favorecen el desarrollo de la malaria. Los mapas auto-organizados de Kohonen fueron utilizados para el análisis de las características más significativas en la clasificación. Los clusters detectados fueron comparados con la clasificación geográfica de las casas, encontrando, que los mapas auto-organizados de Kohonen clasifican las casas por las condiciones ambientales propicias para el desarrollo del mosquito más que por la clasificación administrativa de la ciudad.

Resumo A Metodologia de Clustering foi usada para agrupar três bairros em Quibdo, Colômbia, levando em consideração fatores que favorecem o desenvolvimento da malária. Mapas auto-organizados de Kohonen foram utilizados para a análise das características mais significativas no agrupamento. Os Clusters detectados foram comparados com o agrupamento geo-gráfico de casas, mostrando que os mapas auto-organizados de Kohonen agrupam as casas pelas condições ambientais favoráveis ao desenvolvimento do mosquito e não pelo agrupamento administrativo da cidade.

Abstract Clustering methodology was used to group three neighborhoods in Quibdo taking into account factors that favor the development of malaria. The Kohonen self-organizing maps were used for the analysis of the most significant features in the standings. The detected clusters were compared with the geographical classification of houses, finding that the Kohonen self-organizing maps households classified by environmental conditions conducive to development rather than the administrative classification of the city.