網路輿情與地方施政滿意度指標的觀測 以新竹市長林智堅為例

鉅量資料分析在企業界的應用日漸成熟，是否可以對政治人物也有一套固定的評斷方式，因此本研究選定以 2014 年選上的林智堅市長作為研究對象，透過網路輿情分析工具 OpView 進行分析，以新竹市長林智堅在公開社群媒體當中的網路聲量作為觀察的主體，透過新聞媒體的關鍵字與大數據工具所檢索到的所有有關之文本，進行初步的分析。此類型研究的固定模式如下：設定觀測對象、議題的關鍵字後，可調整觀測周期，使用到的指標有三：網路聲量、網路情緒與討論之詳細議題。在透過此種輿情分析工具，本文設定觀測步驟：先了解網路聲量的大小，再依據網路聲量大的時段觀測網民討論情緒，最後搭配對應文章提取該時段討論的主要內容，即時回報並反映給執政團隊協助決策。以此種分類標準可以分出4種類別：優勢、觀察、警戒、危機共四種。可以更容易進行識別與後續回應流程排序。但此種分析方式雖能夠大量節省搜尋時間，仍需搭配專業團隊進行解讀，並受限於技術或人為操作的因素，不容易與傳統施政滿意度搭配。而本次林智堅勝選的原因，透過輿情工具分析的結果來看，整體網路討論整體情緒在2015年甫上任時較差；在 16 與 17 年透過「有感施政」上先取得了不少分數成為連任的最大武器；但2018年的選舉年中，評論甚至比起15年要差。但依靠媒體與政黨協助在「政治合體」上也對應了年輕世代的喜好，奠定勝選基礎。最後，本文建議在實務層面上，本研究建議可以朝幾個方向持續研究：第一，網路輿情探勘技術(網路輿情與議題管理）；第二，挑選適合自己的輿情探勘。本研究建議，政治人物與政府機關部門，在管理與自身相關的網路議題時，可以仰仗具有媒體公關與行銷的專業人士，並應更重視網路議題管理的重要性與內涵，依照執政者治理需求作為核心的目標，將網路輿情分析技術的導入執政團隊，才可能助於提升工作效率與效能，進而回應民意。

關鍵字

鉅量資料分析；網路輿情；地方施政滿意度；網路聲量； OpView

並列摘要

Big data is becoming more mature in the business world. But is it also a useful tool to be applied to the political world? This study aimed to collect and analyze the public opinions about Mayor Chih-Chien Lin, who was elected as the mayor of Hsinchu City in 2014. With OpView, public opinions related to Mayor Lin on social media were identified and preliminarily analyzed. For this study and other similar studies, three regular steps are built, inclusive of target setting, keyword setting, and period adjusting. There are also three indices used for this type of studies, including volume, public emotion and detailed discussion. This study has set up a standard procedure of observation: first, monitor the volume of public opinions; second, identify public emotions while the volume accumulates to some degree, and then, fetch current online posts to the team for detailed analyses. There are four categories of signals defined, inclusive of advantage, pending, alert and crisis which are helpful for further issue identification and processing. Although this technique is good at collecting a massive database and keeping temporal costs down, it is still necessary to be complemented with professional consulting and constrained by manual operations. As for the winning factors of Mayor Lin's 2018 election, although overall public satisfaction was poorer than 2015, both policies of "non-trivial administration" and "political integration" in 2016 and 2017 had led Mayor Lin to the last victory. This study also proposes two suggestions about the technique of public opinion mining and the pertinence of the issue to be mined. It is suggested that political figures could recruit professional staff of media public relationship and marketing to manage their online issues and emphasize the importance and quality of such issues. In addition, they could base on the core policy to introduce public opinion analysis techniques to improve administrative efficiency and effectiveness, and public satisfaction.