利用政府開放性資料建構台灣線上互動式疾病死因地圖

Using government open data to construct a Taiwan online interactive map of disease causes of death

目標：死因統計是評量健康城市、探討地區疾病負擔的重要指標，隨著開放資料的浪潮，鄉鎮死因統計數據已開放給全民使用，本研究建構線上互動式死因地圖，期能讓各界更方便探索長時期鄉鎮死因的時空變化。方法：使用輕量程式語言JavaScript進行網站功能的開發，死因資料來源為政府開放平台中1991至2013年全死因統計與癌症死因統計，人口統計資料來自內政部，用以計算年齡標準化死亡率，空間統計功能使用Local Moran's I方法進行死亡率空間聚集的偵測。結果：在線上地圖中(http://mortality.geohealth.tw)使用者可以探索縣市或鄉鎮、不同年度、性別、死因的空間分佈，透過動畫方式可一次瀏覽多年度死因資料，並可透過空間統計功能找出特定死因聚集的熱區。若選定某一特定縣市或鄉鎮，互動式統計圖將會顯示該縣市或鄉鎮死亡率位於全國的排名、分性別死亡率、時間趨勢圖，亦可顯示各縣市或鄉鎮的十大死因。結論：藉此網站同時可以查詢死因統計在各地區的時空變化與各種死因的組成方式，將有利於衛生政策制訂與探討健康不平等議題。

Objectives: Statistics on causes of death are an important indicator to measure healthy cities and their disease burdens. With the trend of increasingly accessible government data, township statistics on causes of deaths have already been made available to the public. This study constructs an online interactive map of causes of death to facilitate exploring the long-term spatiotemporal trends. Methods: We used a light programming language - JavaScript - to develop the website. The data of causes of death including all causes and cancer-specific ones during 1991- 2013 are from the platform of publicly accessible government data. The population data are from the Ministry of Interior. We utilized them to compute age-adjusted mortality rates. The spatial statistic, Local Moran's I, was used to detect spatial clusters of mortality. Results: In the online map (http://mortality.geohealth.tw/), users can explore the spatial distribution of mortality at the city or township level, for different years, genders, and causes. Through an animation function, we can compare the same cause of death in many years, and we can also use the spatial statistic to identify the hotspots of some causes. If we select one city or township, the interactive statistical chart displays the ranking of mortality among all cities or townships, gender-specific mortality, temporal trends and the top 10 leading causes of death. Conclusions: With this website, users can simultaneously elucidate spatio-temporal changes of causes of death and major components of causes at the city or township level, which can be beneficial for formulating health policy and studying the issues of health disparity.