Analysing real world data streams with spatio-temporal correlations: Entropy vs. Pearson correlation

Reflections on Evolving Urban Correlations for Connected Smart Cities

“…Creswell (2018) indicates that correlation is also used to "predict an outcome" noting that types of correlational designs include explanatory design and prediction designs (2018). Correlation coefficients are described by Bermudez-Edo et al (2018) as "tools to measure the dependency of two or more variables" such as the Pearson correlation coefficient and that of mutual information. Kaethler (Kaethler and Golchehr, 2017) problematizes "the correlation-based approach" to urban areas in the form of a lament, arguing that, "data mining will disorient the emergence and retention of urban narratives."…”

“…Drawing on the research work of Gehl (1987) and his partner, Ingrid Mundt, Covatta (2018) notes that "data collection regarding citizens' movement and behavior become precious information for governments and municipalities" in support of "participatory design proposals" that are "based on the small scale intervention of people's behavior." Bermudez-Edo et al (2018) provide an analysis of "real world data streams" using "spatio-temporal correlations" in smart cities as a "proposed method" to "analyze relations in IoT." Bermudez-Edo et al (2018) argue that, "one of the main starting points to analyse these big amounts of data is correlation detection" claiming that their "analysis contributes to optimize the use of the available infrastructure" in order "to better plan the future urban deployments."…”

“…Bermudez-Edo et al (2018) provide an analysis of "real world data streams" using "spatio-temporal correlations" in smart cities as a "proposed method" to "analyze relations in IoT." Bermudez-Edo et al (2018) argue that, "one of the main starting points to analyse these big amounts of data is correlation detection" claiming that their "analysis contributes to optimize the use of the available infrastructure" in order "to better plan the future urban deployments." Focusing on "road infrastructure" and traffic flow, the work of Bermudez-Edo et al (2018) "demonstrated that for both, synthetic and real data, International Conferences ICT, Society, and Human Beings 2019; Connected Smart Cities 2019; and Web Based Communities and Social Media 2019 mutual information can discover dependencies in more general cases of distribution of data" while "Pearson correlation can discover linear distribution of data" concluding that "sample size is important" in that "it influences the correlation methods.…”

“…Going forward, the use of correlation in smart environments provides an opportunity or, in the words of researchers such as Bermudez-Edo et al (2018), a "starting point" for the analysis of relationships and for "the narratives we create" (Kaethler and Golchehr, 2017) in urban spaces. Indeed, considering the work of Quercia et al (2017) where correlation is used to "make visible" what is not visible, this would seem to support the relevance of correlation for smart city-related research where the characteristic of invisibility is said to be a particular challenge (McKenna, 2019), offering possibly novel analytic methodologies and approaches with implications for urban theory.…”

“…Using multiple methods of data collection including survey and in-depth interviews, a combination of content analysis and explanatory correlational design is used in the analysis of data. This work is motivated by the need for approaches to "make visible what is normally not" (Quercia, Séaghdha, and Crowcroft, 2017) in smart cities while moving beyond the static to the dynamic in accommodating the "connected" and continuous nature of activities and interactions in technology-pervasive urban environments (Bermudez-Edo, Barnaghi, and Moessner, 2018). As such, this paper uses a combination of case study and correlational statistical research to explore patterns and relationships in smart cities.…”