因為普及性、規模性及即時性等三大特性，讓社群媒體被視為極具潛力的公眾事件監控與偵測新載具。雖然目前已有不少研究，利用社群資料及社群聆聽技術，針對流行病監控與重大事件觀測進行實證測試，但類似作法與技術，是否可適用於地區性公眾事件的監控與偵測，則尚未有具體的討論。
因此，為呼應此一研究缺口，本研究鎖定地區性青少年犯罪議題作為關注焦點，透過實驗設計方式，分析應用社群媒體進行地區治安事件偵測之可行性，同時探討關鍵字設定、資料源選擇，以及涵蓋資料時程等執行策略。本研究以2013年至2014年臺灣社群資料與網路資料為標的，透過OpView及各階段實驗設定之關鍵字，進行分析。研究結果發現，受限地區性議題關鍵字相對模糊且資料量相對過小等因素，從社群資料中篩選出可供實務應用的數量不如預期，但是，如和既有做法相比，基於成本效益考量，仍具應用價值。
The popularity and scale of social media have resulted in a growing need (particu-larly at the governmental level) for a method of extracting useful information from sub-stantial amounts of data. Generally, recent studies have convincingly claimed that using social media can facilitate monitoring and managing public issues. Applications for ep-idemiological surveillance (e.g., influenza epidemics and H1N1 outbreaks) and moni-toring for the broader concept of disaster management have been proven as appropriate methods. However, less attention has been focused on the applicability of current tech-niques in detecting and tracking regional disturbances.
To respond to this need, this study examined the feasibility of detecting and track-ing regional disturbances by using information retrieved from multiple social media channels in Taiwan. Public messages posted on Facebook, PTT, blogs, and bulletin boards from September 2013 to August 2014 were used as our data set. On-site field tests on four issues were conducted using proven policies to examine the feasibility of the method and the quality of the study outcomes. The determined features of regional disturbances indicated that the current keyword- and message-search strategy employed in epidemiological surveillance and disaster management might not perform equally ef-fectively in the context of regional disturbance management. However, this method still warrants application from a cost–benefit perspective. Moreover, strategies for selecting keywords and media sources are discussed and suggested, according to the results of the four-stage experiment.
