Identifying characteristics of online civil complaints by text mining and their influence on the crash risk

Abstract

The researches on the civil complaints are not actively performed in the field of urban planning. Previous attempts have focused on the administrative aspects of the E-government or the exploration on the types of civil complaints registered on the specific events. This study aims to further exploration on the online civil complaints and examine the availability of using the complaints as perception data in crash risk research within the context of the Seoul Metropolitan area. The online civil complaints which are registered on Saeol system are used to analyze the characteristics of citizen’s perception on urban problems. Specifically, civil complaints which contain the crash risk contents are extracted to be analyzed as the crash risk indicating data to examine the influence to the actual crash frequency. Applying the text mining technique and network analysis, we qualitatively explored how people recognize the urban problems and developed the method to extract civil complaints which indicate crash risk. Furthermore, we examined whether there exists significant relationship between actual crash and the crash risk complaints as risk perception data. The findings suggest that 1) citizens tend to recognize the urban problems simply in a fragmentary way, 2) there are several types of civil complaints that are constantly being raised, 3) focusing on the safety related civil complaints, we could extract the crash risk complaints, and 4) these crash risk complaints show the significant relationship with the actual crash frequency. This study identifies the characteristics of the online civil complaints with the text mining technique and examine the risk complaints as risk perception data. In this regard, the findings are expected to help urban planners to understand the citizen’s common recognition on the urban problems and develop the appropriate policies for conflict reduction.