個人的理想城市

Abstract

摘 要 每個人對於衡量理想的居住地區準則不同，但現有的調查僅考慮各地區的主客觀因素提供一個綜合性指標，而並未考量到個人化的差異。本研究欲建構一套個人化決策系統，可供使用者選擇自身關心的衡量項目，經由政府開放平台的客觀數據，並以縝密的統計分析方法，最終以視覺化地圖的呈現個人的理想城市排名。 本研究利用R軟體的Shiny套件建構出互動式頁面，此程式須包含前端的互動頁面與後端的統計分析兩大部分，首先，使用者於前端勾選欲關注的細項變數後，後端將以收集的資料以勾選的變數進行主成分分析，並計算出各縣市的理想總分與排名，再依設定好的理想區段將縣市分為三等級，最終將各等級以深淺不同的顏色代表，並以Leaflet套件將顏色繪製於臺灣地圖上，並傳送至前端的頁面中顯示。 我們藉由本次專題研究，發揮於統計學系學習的知識，建構了一個可供個人需求來進行統計分析的系統，利用簡單的互動式操作，即可讓使用者清楚的解讀複雜的統計分析的結果。

Abstract

The criteria to identify ideal countries are different for each person. However, the reports in general most liveable survey did not evaluate the diversity. In this research, we construct a personalized decision system to find my ideal city. Shiny package in R program is used to build the interactive web page. Two parts are essential, including frontend visualization and backend data analysis. Based on the items that user selected, data retrieve from the government open data is utilized to estimate the scores for each city via principal components analysis. The ranks for all cities transformed from scores are shown with colors. The Leaflet package is adopted to display the analytic results on the Taiwan map. In this project, all techniques learned from college are used such as programming and statistical theory. We provide a interactive personalized system to make the statistical results easy to understand.