Tracking Citizen Appeals For Public Utility Services With Visual Text Analytics

With the rapid expansion of city,some infrastructures and public utility services could not keep pace of city development.Such as infrastructure's aging phenomenon and utility service divisions' low efficiency has been exposed and challenged citizens' normal life quality.C itizens are the major utility services object and citizen appeals for public utility services provide a pregnant insight into the weakness during city development.Based on the citizen appeals analysis,infrastructure failures can be found timely and failure reasons can be analyzed.Our research proposed a series of techniques and algorithms to assist city manager in figuring out citizen's requirements and current situation of infrastructures,analyzing anomalies' temporal-spatial distribution patterns and resolve efficiency.However,citizen appeals are complex,having multi-dimensional and multivariate natures,and containing spatial and temporal components in addition to the content theme information.Considering the increasing volumes and oral nature of citizen appeals,it is hard to specify exploratory problems and visual representations with unstructured text content.The majority information of citizen appeals is delivered by the text content.So we learned the related work about words segmentation techniques,topic modeling methods,and relationship analyzing model,as well as the visual text analytics.To analyze these infrastructure weakness,we propose a visual text analytics process based on the main tasks of utility service management.Firstly,we pre-processed citizen appeals,including extracting location information from appeal content,analyzing address unit through IKAnalyzer techniques,translating location information in text to longitude and latitude,and utilizing IC TCLAS techniques to do the word segmentation.Secondly,we extracted topic and build the topic model,including removing stop words,getting text characteristic,and using LDA+Gibbs methods to build the topic model.And we provide a set of tools and interaction techniques from spatial,temporal and semantic perspectives to do the visual analysis.Through the analysis of our experiment,we presented the findings to administrators in related division.They confirmed our findings and explained the reason led to that phenomenon.They were satisfied with our discoveries and thought usefulness of our visual analysis method and our tools.Our research can help administrators extract the information behind the citizen appeal and better know the current situation of city public utility service.Meanwhile,it enable administrators to analyze the evolution of spatial and temporal pattern and explore the relationship between topics,as well as detect the critical events.O ur approach enable administrators grasp the citizen appeals and track the public utility services anomalies to make informed decisions.