Browsing by Author "Gubareva, Regina"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
- Big data trends in the analysis of city resourcesPublication . Gubareva, Regina; Lopes, Rui PedroThe operation and management of a municipality generate large amounts of complex data, enclosing information that is not easy to infer or extract. Their analysis is challenging and requires specialized approaches and tools, usually based on statistical techniques or on machine learning and artificial intelligence algorithms. These Big Data is often created by combining many data sources that correspond to different operational groups in the city, such as transport, energy consumption, water consumption, maintenance, and many others. Each group exhibits unique characteristics that are usually not shared by others. This paper provides a detailed systematic literature review on applying different algorithms to urban data processing. The study aims to figure out how this kind of information was collected, stored, pre-processed, and analyzed, to compare various methods, and to select feasible solutions for further research. The review finds that clustering, classification, correlation, anomaly detection, and prediction algorithms are frequently used. Moreover, the interpretation of relevant and available research results is presented.
- Extracting temporal patterns from smart city dataPublication . Gubareva, Regina; Lopes, Rui Pedro; Burnakulova, G.S.In the modern world data and information become a powerful instrument of management, business, safety, medicine and others. The most fashionable sciences are the sciences which allow us to extract valuable knowledge from big volumes of information. Novel data processing techniques remains a trend for the last five years, in a way that continues to provide interesting results. This paper investigates the algorithms and approaches for processing smart city data, in particular, water consumption data for the city of Bragança, Portugal. Data from the last seven years was processed according to a rigorous methodology, that includes five stages: cleaning, preparation, exploratory analysis, identification of patterns and critical interpretation of the results. After understanding the data and choosing the best algorithms, a web-based data visualizing tools was developed, providing dashboards to geospatial data representation, useful in the decision making of municipalities.
- Literature review on the smart city resources analysis with big data methodologiesPublication . Gubareva, Regina; Lopes, Rui PedroThis article provides a systematic literature review on applying different algorithms to municipal data processing, aiming to understand how the data were collected, stored, pre-processed, and analyzed, to compare various methods, and to select feasible solutions for further research. Several algorithms and data types are considered, finding that clustering, classification, correlation, anomaly detection, and prediction algorithms are frequently used. As expected, the data is of several types, ranging from sensor data to images. It is a considerable challenge, although several algorithms work very well, such as Long Short-Term Memory (LSTM) for timeseries prediction and classification.