應用於大規模人群移動之分析、模擬與視覺化之研究---子計畫四：人群移動網路視覺化

Abstract

近年來，由於大型伺服器愈來愈普級，資料產生的速度愈來愈快。然而這些資料若沒有好好地利用，它們就跟垃圾沒什麼兩樣，只占著空間卻沒有任何用處。在本計畫中，我們將透過視覺方法來分析、內插、及歸納人群移動的資料。一般來說，這些資料不但具有動態及雜亂的特性，也可能來自不同背景如車輛偵測器、捷運站、天氣偵測器、及智慧型手機。我們認為，視覺化方法可以歸納來自不同背景的資料，協助資料探勘與分析，並將訊息傳遞給使用者，是一個非常具有潛力的技術。不同於傳統的資料探勘，本計劃將透過人類對視覺資料強大的理解能力，經由不同的角度來觀察及探索資料，以期發現全自動的資料探勘技術所無法發現的珍貴訊息。因此，我們主要的研究任務為演算法的設計與評估，其中包含投影空間、視覺傳達、互動式的資料檢視平台。我們相信這種實用的視覺化設計，不僅可以幫助研究人員充份地瞭解人群的移動行為，建立高精確度的資料模型，用於日後的模擬及預測，同時也能提供政府官員從歷史資料中找到關鍵內容，協助他們提出施政方針與做出正確的決策。

The popularity of high-end servers nowadays leads to the availability of vast amounts of data. These data, however, would become trash if they are not used. We propose to develop visual means for analyzing, interpreting, and summarizing the human mobility data, which are typically dynamic and unstructured. These data are also of great diversity, coming from vehicle detectors, metro stations, weather detectors, and mobile phones. We believe that visualization is the most effective way to summarize information from different sources and convey it to the users on the move. Unlike traditional data mining techniques, our method lets users observe and explore data from different viewpoints through an interactive interface and discover important messages that are unavailable from fully automatic methods. As a result, our primary research tasks include the design and evaluation of projection spaces, visualization representations and algorithms, and accompanying interaction techniques for viewing and interrogating different aspects of the data. We believe that our effective visualization designs can help researchers better understand the human mobility so as to derive high precision models for simulation and prediction. In addition, these visualization designs also provide government new insights into the human mobility for developing better transportation facilities and making crucial decisions.