A smart city qualitative data analysis model: Participatory crowdsourcing of public safety reports in South Africa

Abstract

Increased urbanization against the backdrop of limited resources complicates city planning and the management of functions, including public safety. The smart city concept can help, but most previous smart city systems have focused on utilizing automated sensors and analyzing quantitative data. In developing nations, limited resources make using the mobile phone to enable the crowdsourcing of qualitative public safety reports from the public a more viable option. However, there is no best practice for analyzing such citizen reports for a smart city in a developing nation. Given the rise of megacities in developing nations, many of which struggle to provide access to vital resources, this study developed and tested a model for guiding the analysis of unstructured natural language texts instead of traditional sensory data. In the study, citizens engaged with the project and 663 usable reports were received. Following a design science approach, the model was developed through an extensive review of related literature, and assessed and refined by observing the associated model prototype. This study emphasizes that a city-specific ontology needs to be developed and that natural language processing should be its focus, specifically within the larger context of our smart city qualitative data analysis (SCQDA) model. Together, these aspects enable this study to contribute practically, as we prove that cities in developing nations can improve the lives of their citizens using that which is already at their disposal instead of specialized (and often expensive) sensory networks.