時間序列之地理資料分析系統

隨著科技時代的進步，網際網路的蓬勃發展，擁有一台手機便能透過網路得到來自世界各地資訊，而大量資訊的更新、變化速度既迅速又頻繁。現今人們被過量資訊疲勞轟炸，此情況下，較難以簡單統計視覺化圖表呈現龐大的數據量，本研究將基於多維度資料的方式進行研究。本研究目標為時間序列之地理資料，並建立一套探索式分析系統。透過原始資料矩陣與計算兩兩樣本間(國家與國家)的關係矩陣，利用降維方法將其結果結合雙標圖與HSV色彩空間，讓樣本與樣本間的關係能透過色階為地圖上色。除了單一地圖顯示高維度資料外，也利用時間序列分析方法和地圖結合應用，以歷史資料預測未來趨勢，並由視覺化圖表顯示。因此，將得到綜觀的世界地圖，和微觀的各個樣本的時間序列資料呈現。並將其結果組成儀表板的系統，進行探索式的資料分析，讓用戶能透過視覺化清楚得知資訊。

關鍵字

高維度視覺化；時間數列；系統

並列摘要

With the progress of the era of science and technology, the vigorous development of the Internet, with a mobile phone can get information from all over the world through the network, and a large amount of information is updated and changed quickly and frequently. Nowadays, people are bombarded by excessive information fatigue, in this case, it is more difficult to simply visualize the huge amount of data in the statistical visualization chart, and this study will be based on multi-dimensional data. The goal of this study is to use geographic data analysis of time series to display its results on the interface. Through the original data matrix and the calculation of the relationship matrix between the two samples (country and country), the dimensionality reduction method is used to combine the results with the double map, and through the HSV color space, the relationship between the sample and the sample can be colored with color scale for the map. In addition to a single map displaying high-dimensional data, time series methods can be combined with maps to predict future trends with historical data and displayed by visual charts. Thus, you will get an overview of the world map, and microscopic presentation of time series data from each sample. The results are formed into a dashboard system that performs exploratory data analysis so that users can clearly understand the information through visualization.