Assessing public opinion using self-organizing maps. Lessons from urban planning in Romania

Public consultation is an essential component of the planning process to achieve sustainable development. However, there is still a lack of knowledge related to public engagement in consultation processes on urban planning matters. Thus, the aim of this study was to understand public needs in terms of planning urban space by exploring the content of online comments made by the public. We used, as a case study, urban planning in Bucharest, Romania. We assessed 1879 online comments submitted between 2008 and 2020 through a municipality-set online platform, to 220 out of the 1148 zonal urban plans issued during that period. The self-organizing map (SOM) method, an unsupervised artificial neural network, was used to visualize and cluster the online comments based on their content. Using this method, 13 clusters were found to show the general landscape of topics in online comments. These revealed three main and frequently discussed issues among the public: urban green spaces (i.e., saving or developing new green spaces), protection of cultural heritage (i.e., preserving the cultural heritage of the city) and administrative issues (i.e., lack of confidence in the decision-making and permit granting process). In terms of dynamics over the analyzed time period, the topics evolved from esthetic and cultural concerns to issues relating to administrative shortcomings and lack of urban green spaces. Furthermore, we investigated the relationship between plan characteristics and comments and found that the objective of the plan and its year of approval have a positive influence on the presence and number of comments. These results are important to produce higher quality urban projects that are better aligned with the public needs. We further suggest the need to support a sustainable urban spatial development by incorporating the real public needs.