劉濱 劉增杰 劉宇 李子文 陳莉 孫中賢 王瑩 張一輝 趙佳盛 張紅斌 劉青
摘 要:數(shù)據(jù)可視化對于從海量數(shù)據(jù)中發(fā)現(xiàn)規(guī)律、增強(qiáng)數(shù)據(jù)表現(xiàn)、提升交互效率具有重要作用。目前,數(shù)據(jù)可視化的概念及相關(guān)研究領(lǐng)域不斷擴(kuò)展,就數(shù)據(jù)類型而言,可視化研究逐漸聚焦于多維數(shù)據(jù)、時序數(shù)據(jù)、網(wǎng)絡(luò)數(shù)據(jù)和層次化數(shù)據(jù)等領(lǐng)域。通過對中國知網(wǎng)(CNKI)中外文文獻(xiàn)進(jìn)行分析可知:2014年、2015年是數(shù)據(jù)可視化領(lǐng)域研究熱度升級、理論成果大量產(chǎn)出的“里程碑”式年份;中國大數(shù)據(jù)領(lǐng)域研究熱潮形成后,數(shù)據(jù)可視化是迅速發(fā)展的一個重要支撐領(lǐng)域;國內(nèi)外數(shù)據(jù)可視化領(lǐng)域的研究,在時間上基本同步,而武漢大學(xué)、浙江大學(xué)、北京郵電大學(xué)、國防科技大學(xué)、電子科技大學(xué)等都是在該領(lǐng)域研究活躍度較高的國內(nèi)高校。
要獲得良好的視覺效果,幫助用戶降低理解難度,高效分析數(shù)據(jù)和洞悉價(jià)值,通常還需要注意色彩與語義、突出核心數(shù)據(jù)、防止數(shù)據(jù)過載、防止思維過度發(fā)散等技術(shù)要點(diǎn)?,F(xiàn)有的數(shù)據(jù)可視化技術(shù)主要分為基于幾何技術(shù)、基于圖標(biāo)技術(shù)、基于降維技術(shù)、面向像素技術(shù)、基于時間序列技術(shù)、基于網(wǎng)絡(luò)數(shù)據(jù)技術(shù)的數(shù)據(jù)可視化方法,以及層次可視化技術(shù)和分布技術(shù)等。基于幾何技術(shù)的可視化方法,包括平行坐標(biāo)、散點(diǎn)圖矩陣、Andrews曲線等?;谧鴺?biāo)的可視化方法,可以清晰展示變量間的關(guān)系,但受限于屏幕尺寸,當(dāng)數(shù)據(jù)維度超過3個時,難以直觀顯示全部維度,需要結(jié)合人機(jī)交互技術(shù)進(jìn)行展示,適用于表達(dá)不同維度之間的相關(guān)關(guān)系,比如學(xué)生學(xué)習(xí)行為之間的關(guān)聯(lián)關(guān)系等。基于圖標(biāo)的可視化方法,主要包括星繪法和Chernoff面法,以幾何圖形作為圖標(biāo)刻畫多維數(shù)據(jù),直觀反映出圖標(biāo)各個維度所表示的意義,適用于工作完成情況、激勵工作進(jìn)度概覽等?;诮稻S技術(shù)的可視化方法,根據(jù)維度屬性確定點(diǎn)的坐標(biāo),在保持?jǐn)?shù)據(jù)關(guān)系不變的前提下映射到低維可視空間中,主要涉及主成分分析、自組織映射、等距映射等。基于時間序列的可視化方法,是一種顯示數(shù)據(jù)間相互關(guān)系和影響程度的可視化方法,主要包含線形圖、堆積圖、地平線圖等,隨著時間發(fā)展采集相應(yīng)數(shù)據(jù),并利用上述3類可視化方法進(jìn)行呈現(xiàn),適用于表示信息數(shù)據(jù)流動和變化狀態(tài),如不同時間段成績流向趨勢分布、主題概念的變遷等?;诰W(wǎng)絡(luò)數(shù)據(jù)的可視化方法,核心是自動布局算法,通過自動布局與計(jì)算繪制成網(wǎng)狀結(jié)構(gòu)圖形,主要有力導(dǎo)向布局、圓形布局、網(wǎng)格布局等,常用來表示大規(guī)模社交網(wǎng)絡(luò)結(jié)構(gòu),適用于活躍度分析、引文關(guān)系展現(xiàn)等。層次可視化技術(shù),主要包括節(jié)點(diǎn)鏈接、空間填充、混合方法等,通過繪制不同形狀的節(jié)點(diǎn)和包圍框來表示層次結(jié)構(gòu)的數(shù)據(jù),適用于表示群組成員間交互關(guān)系的發(fā)現(xiàn)和挖掘,如在線協(xié)作員工之間的交互。
基于CNKI,通過對數(shù)據(jù)可視化研究情況的分析,提出數(shù)據(jù)可視化研究過程中的注意點(diǎn),指出數(shù)據(jù)可視化需要重點(diǎn)考慮色彩的匹配,在色彩與數(shù)據(jù)內(nèi)容的重要度之間建立關(guān)聯(lián);可視化方案應(yīng)在滿足業(yè)務(wù)需求的基礎(chǔ)上以業(yè)務(wù)邏輯為依據(jù),合理組合與應(yīng)用相關(guān)可視化技術(shù);統(tǒng)一的可視化風(fēng)格有助于提升人們理解數(shù)據(jù)的連貫性、一致性和效率,兼顧用戶的審美要求,在風(fēng)格與色彩之間建立合理的匹配關(guān)系;數(shù)據(jù)可視化應(yīng)以實(shí)用、合理、高效地表現(xiàn)關(guān)鍵過程、關(guān)鍵目標(biāo)、關(guān)鍵結(jié)果為主要面向。此外,對可視化應(yīng)用實(shí)例Echarts展開綜述,包括Echarts 交互組件(markPoint和markLine標(biāo)注點(diǎn)組件、dataZoom區(qū)域組件、圖例交互組件)在可視化中的應(yīng)用,以及動態(tài)數(shù)據(jù)繪制等。最后,對可視化存在的挑戰(zhàn)以及未來研究方向進(jìn)行了分析和展望,指出虛擬現(xiàn)實(shí)、可視化系統(tǒng)和數(shù)據(jù)分析是可視化未來的研究方向,其應(yīng)用熱點(diǎn)領(lǐng)域還包括統(tǒng)計(jì)可視化、新聞可視化、思維可視化、社交網(wǎng)絡(luò)可視化和搜索日志可視化等。
關(guān)鍵詞:計(jì)算機(jī)圖形學(xué);數(shù)據(jù)可視化;多維數(shù)據(jù);時序數(shù)據(jù);網(wǎng)絡(luò)數(shù)據(jù);層次化數(shù)據(jù)
中圖分類號:TP393?? 文獻(xiàn)標(biāo)識碼:A
doi:10.7535/hbkd.2021yx06012
Review of data visualization research
LIU Bin1,2,LIU Zengjie1,2,LIU Yu3 ,LI Ziwen4,CHEN Li5,SUN Zhongxian1,2,WANG Ying1,2,ZHANG Yihui1,2,ZHAO Jiasheng1,2,ZHANG Hongbin6,LIU Qing1,2
(1.School of Economics and Management,Hebei University of Science and Technology,Shijiazhuang,Hebei 050018,China;2.Research Center of Big Data and Social Computing,Hebei University of Science and Technology,Shijiazhuang,Hebei 050018,China;3.Library,Hebei Professional College of Political Science and Law,Shijiazhuang,Hebei 050061,China;4.Hebei Institute of Laser Company Limited,Shijiazhuang,Hebei 050081,China;5.Air Force Early Warning Academy,Wuhan,Hubei 430019,China;6.School of Information Science and Engineering,Hebei University of Science and Techno-logy,Shijiazhuang,Hebei 050018,China)
Abstract:
Data visualization plays an important role in discovering rules from massive data,enhancing data performance and improving interaction efficiency.At present,the concept of data visualization and related research fields are expanding.In terms of data types,the current visualization research gradually focuses on the fields of multidimensional data,time series data,network data and hierarchical data.Through the analysis of Chinese and foreign literature on CNKI,
it can be seen that 2014 and 2015 are "milestone" years in which the research heat in the field of data visualization is upgraded and a large number of theoretical achievements are produced;Data visualization is an important supporting field of rapid development after the formation of the research upsurge in the field of big data in China;The research in the field of data visualization at home and abroad has basically achieved synchronization in time;Wuhan University,Zhejiang University,Beijing University of Posts and telecommunications,University of national defense science and technology and University of Electronic Science and technology research actively in this field in China.In order to obtain good visual effects,help users reduce the difficulty of understanding,efficiently analyze data and insight value,It is usually necessary to pay attention to technical points such as color and semantics,highlighting core data,preventing data overload and preventing excessive divergence of thinking.The existing data visualization technologies are mainly divided into geometry based technology,icon based technology,dimension reduction based technology,pixel oriented technology,time series based technology,network data based technology,hierarchical visualization technology and distribution technology.Visualization methods based on geometric technology,including parallel coordinates,scatter matrix,Andrews curve,etc;The coordinate based visualization method can clearly show the relationship between variables,but limited by the screen size,it is difficult to visually display all dimensions when the data dimensions exceed three.It needs to be displayed in combination with human-computer interaction technology,which is suitable for the correlation between different dimensions,such as the correlation between students' learning behaviors;Icon based visualization method mainly includes star drawing method and Chernoff surface method.Geometric graphics are used as icons to depict multi-dimensional data,which intuitively reflects the visual significance of each work surface.It is suitable for work completion and incentive work progress overview,etc;The visualization method based on dimension reduction technology determines the coordinates of points according to the dimension attributes and maps them to the low-dimensional visual space on the premise of keeping the data relationship unchanged.The dimension reduction technology mainly involves principal component analysis,self-organizing mapping,isometric mapping,etc;The visualization method based on time series is a visualization method to display the relationship and influence degree between data,mainly including linear graph,stacking graph,horizon graph,etc.the corresponding data is collected with the development of time and presented by the above three visualization methods,which is suitable for representing the flow and change state of information data,such as the trend distribution of grades in different time periods and the change of theme concepts,etc;The core of the visualization method based on network data is the automatic layout algorithm,which draws the graph of network structure through automatic layout and calculation.It mainly strongly guides the layout,circular layout and grid layout,etc.It is commonly used to represent the large-scale social network structure,which is suitable for activity analysis,citation relationship,etc;Hierarchical visualization technology mainly includes node connection,space filling and hybrid methods,etc.it represents the data of hierarchical structure by drawing nodes and bounding boxes with different shapes.It is suitable for the discovery and mining of interactive relationships among group members,such as the interaction between online collaborative employees.
Based on the analysis of data visualization CNKI research,this paper puts forward some points for attention in the process of data visualization,and points out that data visualization technology needs to focus on color matching and establish a relationship between color and the importance of data content;The visualization scheme shall reasonably combine and apply relevant visualization technologies based on business logic on the basis of meeting business needs;The unified visualization style helps to improve the coherence,consistency and efficiency of people's understanding of data;At the same time,It also takes into account the aesthetic requirements of users and establishes a reasonable matching relationship between style and color;Data visualization should focus on the practical,reasonable and efficient performance of key processes,key objectives and key results.This paper also summarizes the visualization application example Echarts,including the application of Echarts interactive components (markPoint and markLine annotation point components,datazoom area components,legend interactive components) in visualization,dynamic data rendering and so on.Finally,the challenges and future research directions of visualization are analyzed and prospected,and it is pointed out that virtual reality,visualization system and data analysis are the research directions of visualization in the future.Its application also includes statistical visualization,news visualization,thinking visualization,social network visualization and search log visualization.
Keywords:
computer graphics;data visualization;multidimensional data;time series data;network data;hierarchical data
數(shù)據(jù)可視化,是近年來大數(shù)據(jù)領(lǐng)域各界關(guān)注的熱點(diǎn),屬于人機(jī)交互、圖形學(xué)、圖像學(xué)、統(tǒng)計(jì)分析、地理信息等多種學(xué)科的交叉學(xué)科,綜合數(shù)據(jù)處理、算法設(shè)計(jì)、軟件開發(fā)、人機(jī)交互等多種知識和技能,通過圖像、圖表、動畫等形式展現(xiàn)數(shù)據(jù),詮釋數(shù)據(jù)間的關(guān)系與趨勢,提高閱讀和理解數(shù)據(jù)的效率。就數(shù)據(jù)類型而言,當(dāng)前的可視化研究逐漸聚焦于多維數(shù)據(jù)、時序數(shù)據(jù)、網(wǎng)絡(luò)數(shù)據(jù)和層次化數(shù)據(jù)等領(lǐng)域[1-5]。當(dāng)前,武漢大學(xué)、浙江大學(xué)、北京郵電大學(xué)、國防科技大學(xué)等形成了國內(nèi)該研究領(lǐng)域的“第一梯隊(duì)”。由于在業(yè)務(wù)邏輯、數(shù)據(jù)內(nèi)涵、實(shí)施目標(biāo)、設(shè)計(jì)水平等方面存在差異,導(dǎo)致數(shù)據(jù)可視化應(yīng)用存在著一些誤區(qū)和問題。例如:色彩與語義間的失配、核心數(shù)據(jù)被次要數(shù)據(jù)“淹沒”、數(shù)據(jù)過載、思維過度發(fā)散等。本文提出當(dāng)前可視化研究逐漸聚焦在多維數(shù)據(jù)可視化、時間序列數(shù)據(jù)可視化、網(wǎng)絡(luò)數(shù)據(jù)可視化和層次信息數(shù)據(jù)可視化的觀點(diǎn),介紹了各自的代表性方法;同時,給出數(shù)據(jù)可視化主流工具Echarts的相關(guān)應(yīng)用實(shí)例。
1 基于CNKI的研究情況分析
以“數(shù)據(jù)可視化”為篇名檢索詞,在中國知網(wǎng)(CNKI)學(xué)術(shù)期刊和學(xué)位論文庫中精準(zhǔn)檢索,檢索出文獻(xiàn)11 673篇(中文4 763篇、外文6 910篇)。通過分析可知:相關(guān)中文文獻(xiàn)從2014年的402篇,上升至2015年的975篇,考慮到文章的撰寫和發(fā)表周期,可以說2014年、2015年是數(shù)據(jù)可視化領(lǐng)域研究熱度升級、理論成果大量產(chǎn)出的“里程碑”式年份;大數(shù)據(jù)類相關(guān)中文文獻(xiàn)2012年396篇,2013年2 195篇,2014年4 936篇,直到后期的逐年翻倍,如實(shí)反映了中國大數(shù)據(jù)研究從2012年的逐漸興起,到2013年漸熱,再到2014年3月寫入國務(wù)院《政府工作報(bào)告》后形成熱潮的客觀事實(shí),也表明數(shù)據(jù)可視化是中國大數(shù)據(jù)領(lǐng)域研究熱潮形成后迅速發(fā)展的一個重要支撐領(lǐng)域;相關(guān)外文文獻(xiàn)從2014年的243篇,上升至2015年的659篇。雖然中國知網(wǎng)涵蓋的外文文獻(xiàn)有限,但從發(fā)展趨勢來看,國內(nèi)外數(shù)據(jù)可視化領(lǐng)域的研究,在時間上基本同步。
通過分析在數(shù)據(jù)可視化領(lǐng)域發(fā)表論文的高校可知,雖然國內(nèi)有相當(dāng)多的研究成果在國外期刊發(fā)表,但也可推知,武漢大學(xué)、浙江大學(xué)、北京郵電大學(xué)、國防科技大學(xué)、電子科技大學(xué)等都是在該領(lǐng)域研究活躍度較高的高校。
通過對與數(shù)據(jù)可視化主題關(guān)聯(lián)度較高的前20個學(xué)科進(jìn)行分析可知,計(jì)算機(jī)軟件與應(yīng)用學(xué)科的關(guān)聯(lián)度最高,其次是自然地理學(xué)和測繪學(xué)、新聞與傳媒、圖書情報(bào)與數(shù)字圖書館等,計(jì)算機(jī)學(xué)科既是研究數(shù)據(jù)可視化技術(shù)的主力學(xué)科,也是為其他學(xué)科提供技術(shù)和平臺支持的學(xué)科。因此,數(shù)據(jù)可視化是一個面向應(yīng)用、立足實(shí)際的學(xué)科,其研發(fā)成果已經(jīng)融入到測繪、互聯(lián)網(wǎng)、電力、礦業(yè)、電信、建筑、工業(yè)、海洋、航空航天等國民經(jīng)濟(jì)產(chǎn)業(yè),從現(xiàn)實(shí)需求中發(fā)現(xiàn)問題和解決問題,也是推動該領(lǐng)域快速發(fā)展的重要方式。
2 數(shù)據(jù)可視化在應(yīng)用中的注意點(diǎn)
2.1 注重色彩與語義的聯(lián)系
色彩在數(shù)據(jù)可視化中發(fā)揮著重要作用,是人們讀取數(shù)值、感知趨勢、發(fā)現(xiàn)異常的關(guān)鍵視覺編碼元素。
與光學(xué)物理學(xué)、美學(xué)等領(lǐng)域關(guān)注于色彩自身屬性(如波段、美感等)不同,數(shù)據(jù)可視化中的色彩設(shè)計(jì)以讓用戶高效地理解數(shù)據(jù)、發(fā)現(xiàn)規(guī)律、探索任務(wù)為目標(biāo),側(cè)重于挖掘色彩與數(shù)據(jù)、任務(wù)、設(shè)備等可視化應(yīng)用環(huán)境之間的關(guān)聯(lián)關(guān)系[6]。設(shè)計(jì)者要時刻留心色彩與其用處之間的關(guān)系。例如:個推公司制作的交通熱力圖(見圖1),紅色表示人員密集度較高的情況,綠色或者黃色表示密集度較低的情況。
2.2 突出核心數(shù)據(jù)
一個數(shù)據(jù)可視化作品中的數(shù)據(jù)都是極具針對性的,應(yīng)該基于用戶需求展示相關(guān)數(shù)據(jù)。用戶看可視化作品主要通過數(shù)據(jù)圖表的展示盡早找到核心問題的答案。作品顯示的無關(guān)數(shù)據(jù)越多,找出關(guān)聯(lián)信息越困難,就會極度分散注意力,浪費(fèi)時間。例如,圖2所示的電商年度銷售數(shù)據(jù)可視化圖,應(yīng)該將可視化重點(diǎn)放在銷售額、訂單量、完成率、增長率、重點(diǎn)商品的銷售占比、各平臺銷售占比等銷售數(shù)據(jù)上,而非物流配送、消費(fèi)者性別、買了什么品牌的什么產(chǎn)品這些詳細(xì)信息。優(yōu)秀的數(shù)據(jù)可視化作品應(yīng)該優(yōu)先顯示重要信息,而后是關(guān)聯(lián)信息、可操作性信息,其他內(nèi)容則都應(yīng)該盡可能淡化。
2.3 防止數(shù)據(jù)過載
數(shù)據(jù)可視化可以貫穿情況分析與決策支持的全過程,豐富的數(shù)據(jù)可以為決策者提供更多的觀察維度、更多的可分析層次。但是,豐富不等同于單純的數(shù)據(jù)累加,超過正常觀察能力的可視化數(shù)據(jù)會帶來理解和思考上的障礙,影響最終的決策質(zhì)量。圖3給出的2013年MLS(美國職業(yè)足球大聯(lián)盟)薪酬榜數(shù)據(jù)中,給出了太多的球員和薪酬數(shù)據(jù),形成了數(shù)據(jù)過載,難以達(dá)到讓觀察者迅速看清全局、捕獲重點(diǎn)和關(guān)鍵信息的目標(biāo)。針對此問題,應(yīng)對主題和屬性的個數(shù)進(jìn)行合理限定,確保主流閱讀者能夠舒適、高效地觀察和理解。
2.4 避免思維過度發(fā)散
圖4展示了美國貨物貿(mào)易逆差和工廠雇傭員工數(shù)量的關(guān)系,但卻令人難以理解,其有2個主要問題。第1個問題,貿(mào)易逆差的全部數(shù)據(jù)都是負(fù)值,而工廠雇傭人數(shù)全部是正值。在沒有將2組數(shù)據(jù)歸一化到同一尺度的情況下,
將其組合到一幅圖中表達(dá)是不合適的。這種直白的處理方式導(dǎo)致了第2個問題——2組數(shù)據(jù)沒有共享同一個基線。貿(mào)易逆差的基線是圖頂部左半段的紅線,而右邊尺度的基線又在圖表的底部。其實(shí)將2組數(shù)據(jù)組合在一幅圖中是沒有必要的。在重新設(shè)計(jì)的圖表中,貿(mào)易逆差和工廠雇傭人數(shù)之間的關(guān)系更為清晰,僅僅是多占據(jù)了很小的一點(diǎn)額外空間[7]。針對此類問題,應(yīng)該以“奧卡姆剃刀原理”為指導(dǎo),即如無必要,勿增實(shí)體,在一幅圖中,從維度、標(biāo)注等方面,按照簡單直接的原則進(jìn)行設(shè)計(jì)。
3 基于特征的主要數(shù)據(jù)可視化技術(shù)
作為一門問題和目標(biāo)主要來自于現(xiàn)實(shí)世界的學(xué)科,數(shù)據(jù)可視化在很多領(lǐng)域獲得了研究、應(yīng)用和長足進(jìn)步,在從研究范圍(廣度)、研究精細(xì)化(深度)不斷拓展學(xué)科邊界的過程中,逐漸收斂成為若干熱點(diǎn)領(lǐng)域。本文以數(shù)據(jù)特征來劃分,介紹其中的4類技術(shù)。
3.1 多維數(shù)據(jù)可視化
3.1.1 基于幾何的可視化方法
平行坐標(biāo):使用平行豎直的線來代表不同的維度,在坐標(biāo)軸上描繪多維數(shù)據(jù)的數(shù)值并連接數(shù)軸上的坐標(biāo)點(diǎn),進(jìn)而在二維空間內(nèi)展示多維數(shù)據(jù)[8-11]。在可視化交互方面,有學(xué)者提出了基于平行坐標(biāo)系的多維數(shù)據(jù)可視化方法,以交互方式對數(shù)據(jù)進(jìn)行篩選,通過更改顯示比例優(yōu)化可視化效果[12-13];在可視化應(yīng)用方面,雷君虎等[14]提出了利用PCA主成分分析法對高維度的數(shù)據(jù)進(jìn)行降維,降維后的數(shù)據(jù)通過平行坐標(biāo)系可視化,應(yīng)用于香精香料指紋圖譜數(shù)據(jù);在可視化增強(qiáng)方面,郭翰琦等[15]提出設(shè)置傳遞函數(shù)、對顏色和透明度等可視化特征定義數(shù)值,優(yōu)化圖像效果。
散點(diǎn)圖矩陣:通過二維坐標(biāo)系中的某一組點(diǎn)來展示變量間的關(guān)系,將各個維度數(shù)據(jù)兩兩組合,按規(guī)律排列繪制成散點(diǎn)圖[16],運(yùn)用可視化方法與散點(diǎn)圖矩陣相結(jié)合,加強(qiáng)對多維數(shù)據(jù)效果的顯示。使用散點(diǎn)矩陣圖可以清晰地發(fā)現(xiàn)變量之間的關(guān)系,但受限于屏幕尺寸,當(dāng)數(shù)據(jù)維度超過3個時,難以直觀顯示全部維度,需要結(jié)合人機(jī)交互技術(shù)進(jìn)行展示[17-19]。
Andrews曲線法:通過坐標(biāo)系展示可視化效果,將多維數(shù)據(jù)通過周期函數(shù)反映到坐標(biāo)系曲線中,用戶通過觀察曲線,感知數(shù)據(jù)聚類等情況。
3.1.2 基于圖標(biāo)的可視化方法
基于圖標(biāo)的可視化方法主要是用幾何圖形作為圖標(biāo)對多維數(shù)據(jù)進(jìn)行描繪,圖標(biāo)的特征屬性(硬度、形狀、長短、大小等)體現(xiàn)出信息的維度,利用圖標(biāo)與多維數(shù)據(jù)之間的聯(lián)系反映可視化效果[20]。基于圖標(biāo)的可視化代表性方法有星繪法(見圖5)和Chernoff面法(見圖6)。星繪法通過點(diǎn)到線的方式映射出信息維度,線段長度反映不同維度的數(shù)量值;
Chernoff面法通過對面部形狀、特征等進(jìn)行識別體現(xiàn)信息維度,并繪制臉部圖,直觀觀察信息數(shù)據(jù)。由于Chernoff面法更加有趣高效,有利于識別各個重要特征和要素之間的關(guān)系,所以更多用戶選擇使用Chernoff面法[16]。
3.1.3 基于降維映射的可視化方法
降維映射可視化方法把多維信息數(shù)據(jù)看做是某一維度中的點(diǎn),根據(jù)維度屬性確定點(diǎn)的坐標(biāo),保持信息數(shù)據(jù)間關(guān)系不改變的前提下,將點(diǎn)映射到可視的低維空間中[21]。在降維時選擇性省略掉部分信息數(shù)據(jù),最終在二三維空間中呈現(xiàn)出數(shù)據(jù)集。此類方法主要涉及主成分分析、自組織映射、等距映射等方法。降維映射一般分為線性降維(如主成分分析)和非線性降維(如等距映射),通過特征選擇與提取來實(shí)現(xiàn)。特征選擇通過選擇具有代表性的特征屬性(簡稱優(yōu)勢維)進(jìn)行降維映射,特征提取則是重組多維度屬性來構(gòu)建優(yōu)勢維度,實(shí)現(xiàn)降維映射,這類提取適合沒有代表性特征屬性的信息數(shù)據(jù)集。
3.2 時間序列數(shù)據(jù)可視化
時間序列可視化隨著時間的發(fā)展采集信息數(shù)據(jù),運(yùn)用可視化技術(shù)手段進(jìn)行呈現(xiàn),呈現(xiàn)出的可視化方式主要有3種。一是線形圖(見圖7),通過最開始的點(diǎn)展示不同時間段信息數(shù)據(jù)變化,在可視化過程中信息數(shù)據(jù)呈現(xiàn)較多時間維度,根據(jù)不同維度建立相應(yīng)圖標(biāo)進(jìn)行排列,觀察數(shù)據(jù)的變化;二是堆積圖(見圖8),這類圖主要對所有時間序列進(jìn)行疊加,出現(xiàn)負(fù)數(shù)時,堆積圖無法處理所有的時間序列,極大程度降低了可視化的呈現(xiàn)效果;三是地平線圖(見圖9),隨著時間變化清楚地觀察到信息數(shù)據(jù)的變化率,顏色的深淺表示正向、負(fù)向的變動效果[23-25]。
3.3 網(wǎng)絡(luò)數(shù)據(jù)可視化
網(wǎng)絡(luò)數(shù)據(jù)可視化技術(shù)手段核心是自動布局算法,將信息數(shù)據(jù)通過自動布局、計(jì)算,繪制成網(wǎng)狀結(jié)構(gòu)的圖形。應(yīng)用較廣泛的有3類。力導(dǎo)向布局(見圖10):借助力的概念,連接受力節(jié)點(diǎn)繪制網(wǎng)狀圖,由于互斥力的存在,可以減少節(jié)點(diǎn)間的重疊,適用于描述事物之間的關(guān)系,例如計(jì)算機(jī)網(wǎng)絡(luò)關(guān)系、社交網(wǎng)絡(luò)關(guān)系等各類關(guān)系網(wǎng)絡(luò)情景[26-29]。圓形布局(見圖11):將所有節(jié)點(diǎn)自定義排序,按照順序在圓形上排列出來,快速分析出結(jié)果,受限于屏幕大小,節(jié)點(diǎn)數(shù)量較多時,圓形半徑越來越大,難以直觀顯示全部節(jié)點(diǎn),適用于查找較多關(guān)聯(lián)關(guān)系的節(jié)點(diǎn)場景,例如在圓形布局圖中可明顯分辨出哪些節(jié)點(diǎn)關(guān)聯(lián)關(guān)系較多。網(wǎng)格布局(見圖12):采用網(wǎng)格設(shè)計(jì)方式繪制網(wǎng)格狀信息數(shù)據(jù)網(wǎng)狀圖,適用于分層網(wǎng)絡(luò),利于觀察整體層次[30-35]。
3.4 層次信息數(shù)據(jù)可視化
層次結(jié)構(gòu)常被用來描述具有明顯層次結(jié)構(gòu)的對象,包括圖書館標(biāo)簽、計(jì)算機(jī)層次系統(tǒng)或者面向?qū)ο蟪绦蝾愔g的繼承關(guān)系等[36-40]。層次信息數(shù)據(jù)可視化用到的方法主要包括節(jié)點(diǎn)連接、空間填充、混合方法等。
3.4.1 節(jié)點(diǎn)連接
節(jié)點(diǎn)連接主要繪制不同形狀節(jié)點(diǎn)表示信息數(shù)據(jù)內(nèi)容,節(jié)點(diǎn)之間連線表示數(shù)據(jù)之間的關(guān)系(見圖13)。此類層次代表技術(shù)有空間樹、圓錐樹(見圖14)等。
3.4.2 空間填充
空間填充主要運(yùn)用包圍框表示層次結(jié)構(gòu)信息數(shù)據(jù),上層節(jié)點(diǎn)與下層節(jié)點(diǎn)之間包圍關(guān)系表示信息數(shù)據(jù)間的結(jié)構(gòu)關(guān)系[41]。此類層次代表技術(shù)有樹圖(見圖15)、信息立方體(見圖16)等。
3.4.3 混合方法
混合方法結(jié)合多種可視化技術(shù)優(yōu)點(diǎn),使認(rèn)知行為更加高效[42]。此類方法代表技術(shù)有彈性層次(見圖17)、層次網(wǎng)等。
綜上可知,按照可視化技術(shù)劃分,可視化特征及應(yīng)用場景詳見表1。
4 數(shù)據(jù)可視化應(yīng)用實(shí)例
當(dāng)前涌現(xiàn)出一批專注于數(shù)據(jù)可視化領(lǐng)域的企業(yè),如帆軟、數(shù)字冰雹、蛛網(wǎng)時代、Echarts數(shù)據(jù)可視化等,其以數(shù)據(jù)可視化為基礎(chǔ),根據(jù)用戶需求設(shè)計(jì)不同模式和主題,控制數(shù)據(jù)呈現(xiàn)形式和視覺效果[43-47]。Echarts客戶端構(gòu)建前端界面使用script引入Echarts依賴JS庫。客戶端引入圖表庫流程如圖18所示。
數(shù)據(jù)可視化呈現(xiàn)界面并非靜止不變,還包括用戶交互體驗(yàn)感、使用感受和圖像反饋等。Echarts推出大量交互組件應(yīng)用于可視化中[48-53],成為用戶深入分析了解數(shù)據(jù)的關(guān)鍵手段。
4.1 markPoint和markLine標(biāo)注點(diǎn)組件數(shù)據(jù)中往往呈現(xiàn)最大、最小值,markPoint和markLine組件分別用來表示增加標(biāo)注點(diǎn)和圖表標(biāo)線。如圖19所示,在數(shù)據(jù)可視化銷量類別顯示中,markPoint和markLine標(biāo)注點(diǎn)組件顯示數(shù)據(jù)。
4.2 dataZoom區(qū)域縮放組件
dataZoom區(qū)域縮放組件提供了一種人機(jī)交互能力,主要是為了達(dá)到區(qū)域縮放的效果。通過坐標(biāo)軸的左右平移進(jìn)行縮放,用戶可以了解到數(shù)據(jù)的細(xì)節(jié),總覽整體,去除離散數(shù)據(jù)的影響;還可以觀察個別數(shù)據(jù)的走勢,有針對性地掌握重要數(shù)據(jù)。在可視化系統(tǒng)設(shè)計(jì)中,如果想關(guān)注某個專業(yè)的信息數(shù)據(jù)細(xì)節(jié),則需要加入dataZoom區(qū)域縮放組件進(jìn)行縮放,如圖20所示。
4.3 圖例交互組件
通過點(diǎn)擊圖例標(biāo)記展示對應(yīng)的數(shù)據(jù)系列,幫助用戶排除不必要的數(shù)據(jù)系列,將關(guān)注點(diǎn)放在目標(biāo)數(shù)據(jù)上。如圖21所示,通過點(diǎn)擊圖例進(jìn)行圖例交互,顯示單個圖例所對應(yīng)的系列數(shù)據(jù)。
4.4 動態(tài)數(shù)據(jù)繪制流程
數(shù)據(jù)可視化的圖形繪制中,并不全部是靜態(tài)數(shù)據(jù),動態(tài)數(shù)據(jù)的繪制要通過Ajax技術(shù),從服務(wù)器讀取進(jìn)行動態(tài)加載后呈現(xiàn)出來[54-55],基本流程如圖22所示。
5 研究展望
有研究認(rèn)為,數(shù)據(jù)可視化高頻關(guān)鍵詞為信息可視化、可視化分析、大數(shù)據(jù)和數(shù)據(jù)挖掘,其未來發(fā)展方向?yàn)閰f(xié)同分析和計(jì)算等[56-62]。通過對可視化領(lǐng)域文獻(xiàn)進(jìn)行計(jì)量,根據(jù)關(guān)鍵詞共現(xiàn)圖和時區(qū)圖中的可視化分析、人機(jī)交互和信息可視化等研究熱點(diǎn),結(jié)合關(guān)鍵詞突變圖,筆者對數(shù)據(jù)可視化未來的熱點(diǎn)和發(fā)展趨勢進(jìn)行全景式分析展望,認(rèn)為除了虛擬現(xiàn)實(shí)、可視化系統(tǒng)和數(shù)據(jù)分析外,可視化未來研究方向還包括以下內(nèi)容。
1)自動、智能化 目前數(shù)據(jù)融合越來越復(fù)雜,將工作流自動化和可視化組件結(jié)合,可以有效提高工作效率和質(zhì)量;將智能科學(xué)和可視化結(jié)合,可以利用智能科學(xué)認(rèn)知,彌補(bǔ)人類感知能力和可視化表達(dá)方面的不足,克服復(fù)雜數(shù)據(jù)分析和理解中的難題。
2)協(xié)同可視化 在可視化實(shí)現(xiàn)過程中,需要多團(tuán)隊(duì)協(xié)作完成,創(chuàng)造出協(xié)同可視化的環(huán)境(可視化接口設(shè)計(jì)、可視化協(xié)同平臺開發(fā)、協(xié)同可視化視圖設(shè)計(jì)、工作流管理等),進(jìn)行工作站之間的數(shù)據(jù)資源共享,通過對可視化過程進(jìn)行控制,解決多團(tuán)隊(duì)之間的協(xié)同性問題。
3)應(yīng)用領(lǐng)域拓展化 數(shù)據(jù)可視化技術(shù)已經(jīng)被應(yīng)用到越來越多的領(lǐng)域,既促進(jìn)了各個領(lǐng)域的發(fā)展,也為可視化技術(shù)自身發(fā)展和完善提供了良好環(huán)境。未來可視化的應(yīng)用熱點(diǎn)領(lǐng)域還包括統(tǒng)計(jì)可視化、新聞可視化、思維可視化、社交網(wǎng)絡(luò)可視化和搜索日志可視化等。
數(shù)據(jù)可視化是一個以實(shí)際應(yīng)用為主要面向的人機(jī)交互、圖形學(xué)、圖像學(xué)等多學(xué)科交叉的領(lǐng)域,計(jì)算機(jī)科學(xué)是推動其研究向縱深發(fā)展的主力學(xué)科,自然地理和測繪、新聞與傳媒、圖書情報(bào)、自動化、互聯(lián)網(wǎng)等領(lǐng)域的需求推動了其應(yīng)用領(lǐng)域的不斷拓展。未來還需要對國外研究動態(tài)進(jìn)行更為全面詳細(xì)的梳理,并對數(shù)據(jù)可視化應(yīng)用領(lǐng)域具有代表性的重要案例進(jìn)行更為深入的探討和分析。
參考文獻(xiàn)/References:
[1] DOU Hui,XU Baile,SHEN Furao,et al.V-SOINN:A topology preserving visualization method for multidimensional data[J].Neurocomputing,2021,449:280-289.
[2] EADIE A,VSQUEZ I C,LIANG Xuefeng,et al.Transcriptome network data in larval zebrafish (Danio rerio) following exposure to the phenylpyrazole fipronil[J].Data in Brief,2020,33:106413.
[3] WANG Qiyao,F(xiàn)ARAHAT A,GUPTA C,et al.Deep time series models for scarce data[J].Neurocomputing,2021,456:504-518.
[4] PASPATIS I,TSOHOU A,KOKOLAKIS S.AppAware:A policy visualization model for Mobile applications[J].Information and Com-puter Security,2020,28(1):116-132.
[5] 任磊,杜一,馬帥,等.大數(shù)據(jù)可視分析綜述[J].軟件學(xué)報(bào),2014,25(9):1909-1936.
REN Lei,DU Yi,MA Shuai,et al.Visual analytics towards big data[J].Journal of Software,2014,25(9):1909-1936.
[6] 曾瓊,汪云海,屠長河,等.面向數(shù)據(jù)可視化的色彩計(jì)算[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2020,32(10):1549-1559.
ZENG Qiong,WANG Yunhai,TU Changhe,et al.Color computing in data visualization[J].Journal of Computer-Aided Design & Computer Graphics,2020,32(10):1549-1559.
[7] 帆軟.數(shù)據(jù)可視化的三大問題+七大誤區(qū)[EB/OL].https://zhuanlan.zhihu.com/p/105053580,2021-09-01.
[8] 石凱,聶富強(qiáng),孫峰.多維數(shù)據(jù)判別分析的非參核密度算法研究[J].計(jì)算機(jī)工程與應(yīng)用,2019,55(6):8-12.
SHI Kai,NIE Fuqiang,SUN Feng.Research on algorithm of nonparametric kernel density for discriminant analysis of multidi-mensional data[J].Computer Engineering and Applications,2019,55(6):8-12.
[9] 閆涵,張旭秀,張凈丹.多感知興趣區(qū)域特征融合的圖像識別方法[J].智能系統(tǒng)學(xué)報(bào),2021,16(2):263-270.
YAN Han,ZHANG Xuxiu,ZHANG Jingdan.Image recognition method based on multi-perceptual interest region feature fusion[J].Journal of Intelligent Systems,2021,16(2):263-270.
[10]FRANCIA M,MARCEL P,PERALTA V,et al.Enhancing cubes with models to describe multidimensional data[J/OL].Information Systems Frontiers.[2021-09-01].https://pubmed.ncbi.nlm.nih.gov/34131390/.
[11]PENG Zhen,LIU Wanquan,AN Senjian.Haze pollution causality mining and prediction based on multi-dimensional time series with PS-FCM[J].Information Sciences,2020,523:307-317.
[12]CHEN Yi,CHENG Xianchen,CHEN Hongqian.A multidimensional data visualization method based on parallel coordinates and enhanced ring[C]//Proceedings of 2011 International Conference on Computer Science and Network Technology.[S.l.]:IEEE,2011:2224-2229.
[13]張?jiān)Q,高亞琳,蔣建波,等.一種基于平行坐標(biāo)系的流轉(zhuǎn)數(shù)據(jù)可視化方法[J].計(jì)算機(jī)應(yīng)用與軟件,2018,35(4):55-60.
ZHANG Yuanming,GAO Yalin,JIANG Jianbo,et al.A visualization method for flow data based on parallel coordinates[J].Computer Applications and Software,2018,35(4):55-60.
[14]雷君虎,楊家紅,鐘堅(jiān)成,等.基于PCA和平行坐標(biāo)的高維數(shù)據(jù)可視化[J].計(jì)算機(jī)工程,2011,37(1):48-50.
LEI Junhu,YANG Jiahong,ZHONG Jiancheng,et al.High-dimensional data visualization based on principal component analysis and para-llel coordinate[J].Computer Engineering,2011,37(1):48-50.
[15]郭翰琦,袁曉如.體數(shù)據(jù)可視化傳遞函數(shù)研究[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2012,24(10):1249-1258.
GUO Hanqi,YUAN Xiaoru.Survey on transfer functions in volume visualization[J].Journal of Computer-Aided Design & Computer Graphics,2012,24(10):1249-1258.
[16]楊彥波,劉濱,祁明月.信息可視化研究綜述[J].河北科技大學(xué)學(xué)報(bào),2014,35(1):91-102.
YANG Yanbo,LIU Bin,QI Mingyue.Review of information visualization[J].Journal of Hebei University of Science and Technology,2014,35(1):91-102.
[17]王賦攀,吳亞東,楊文超,等.一種視覺信息融合數(shù)據(jù)手套設(shè)計(jì)研究[J].計(jì)算機(jī)研究與發(fā)展,2018,55(12):2764-2774.
WANG Fupan,WU Yadong,YANG Wenchao,et al.Design and research of data glove based on visual information fusion[J].Journal of Computer Research and Development,2018,55(12):2764-2774.
[18]馮志全,喬宇,馮仕昌,等.多種手勢對應(yīng)同一語義的柔性映射交互算法的研究[J].電子學(xué)報(bào),2019,47(8):1612-1617.
FENG Zhiquan,QIAO Yu,F(xiàn)ENG Shichang,et al.Research on flexible mapping algorithm of multi-gestures to one semantic[J].Acta Electronica Sinica,2019,47(8):1612-1617.
[19]CIFLIKLI C,TARTAN E O.A model for the visualization and analysis of elevator traffic[J].Transportation Planning and Technology,2019,42(8):868-880.
[20]鄭婭峰,趙亞寧,白雪,等.教育大數(shù)據(jù)可視化研究綜述[J].計(jì)算機(jī)科學(xué)與探索,2021,15(3):403-422.
ZHENG Yafeng,ZHAO Yaning,BAI Xue,et al.Survey of big data visualization in education[J].Journal of Frontiers of Computer Science & Technology,2021,15(3):403-422.
[21]翟擎辰,周園春,宋秋成,等.數(shù)據(jù)降維及聚類算法在煙葉相似性分析中的應(yīng)用[J].數(shù)據(jù)與計(jì)算發(fā)展前沿,2021,3(1):112-121.
ZHAI Qingchen,ZHOU Yuanchun,SONG Qiucheng,et al.Application of data dimensionality reduction and clustering algorithm in tobacco leaf similarity analysis[J].Frontier of Data and Computing Development,2021,3(1):112-121.
[22]陳法法,湯寶平,蘇祖強(qiáng).基于等距映射與加權(quán)KNN的旋轉(zhuǎn)機(jī)械故障診斷[J].儀器儀表學(xué)報(bào),2013,34(1):215-220.
CHEN Fafa,TANG Baoping,SU Zuqiang.Rotating machinery fault diagnosis based on isometric mapping and weighted KNN[J].Journal of Instrumentation,2013,34(1):215-220.
[23]CHEN Bingkun,HONG Zhou,CHEN Xiaojun.E-Embed:A time series visualization framework based on earth mover's distance[J].Journal of Visual Languages & Computing,2018,48:110-122.
[24]ALI M,JONES M W,XIE Xianghua,et al.TimeCluster:Dimension reduction applied to temporal data for visual analytics[J].The Visual Computer,2019,35:1013-1026.
[25]SUH A,HAJIJ M,WANG Bei,et al.Persistent homology guided force-directed graph layouts[J].IEEE Transactions on Visualization and Computer Graphics,2020,26(1):697-707.
[26]韓劉.多元異構(gòu)網(wǎng)絡(luò)復(fù)雜多維數(shù)據(jù)可視化方法[J].計(jì)算機(jī)仿真,2020,37(11):299-303.
HAN Liu.Visualization method of complex multidimensional data in multiple heterogeneous networks[J].Computer Simulation,2020,37(11):299-303.
[27]王悅.可視化指導(dǎo)下的圖像領(lǐng)域深度學(xué)習(xí)模型優(yōu)化方案設(shè)計(jì)[D].合肥:中國科學(xué)技術(shù)大學(xué),2020.
WANG Yue.Design of Optimization Scheme for Deep Learning Model in Image Field Under Visual Guidance[D].Hefei:University of Science and Technology of China,2020.
[28]涂樂,陳彬捷,周志光.OD數(shù)據(jù)可視分析綜述[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2021,33(8):1160-1171.
TU Le,CHEN Binjie,ZHOU Zhiguang.A survey on OD data visualization[J].Journal of Computer-Aided Design & Computer Graphics,2021,33(8):1160-1171.
[29]尹彥,張紅斌,劉濱,等.網(wǎng)絡(luò)安全態(tài)勢感知中的威脅情報(bào)技術(shù)[J].河北科技大學(xué)學(xué)報(bào),2021,42(2):195-204.
YIN Yan,ZHANG Hongbin,LIU Bin,et al.Threat intelligence technology in network security situation awareness[J].Journal of Hebei University of Science and Technology,2021,42(2):195-204.
[30]田豐,喻純.自然人機(jī)交互新進(jìn)展專題前言[J].軟件學(xué)報(bào),2019,30(10):2925-2926.
[31]許世健.沉浸式流場可視化交互技術(shù)研究[D].綿陽:西南科技大學(xué),2021.
XU Shijian.Research on Interaction Technology of Immersive Flow Visualization[D].Mianyang:Southwest University of Science and Technology,2021.
[32]李恒,沈華偉,黃蔚,等.地理社會網(wǎng)絡(luò)數(shù)據(jù)可視化分析研究綜述[J].中文信息學(xué)報(bào),2018,32(10):11-18.
LI Heng,SHEN Huawei,HUANG Wei,et al.Research overview of geo-social networks data visualization and analysis[J].Journal of Chinese Information Processing,2018,32(10):11-18.
[33]JO S,PARK B,LEE S,et al.OLGAVis:On-line graph analysis and visualization for bibliographic information network[J].Applied Sciences,2021,11(9):3862.
[34]MCBRIDE M.Data visualization in society[J].Journal of the Australian Library and Information Association,2021,70(1):91-92.
[35]WANG Peng.Visualization research on scientific and technological information of aviation industry in Shaanxi province[J].E3S Web of Conferences,2020,189:03014.
[36]田宇荃,趙迪,陳虎.海綿城市建設(shè)前后的可視化效果分析——以常德市為例[J].智能建筑與智慧城市,2020(6):98-99.
TIAN Yuquan,ZHAO Di,CHEN Hu.Analysis on the visual effect of sponge city before and after construction:Taking Changde city as an example[J].Intelligent Building & City Information,2020(6):98-99.
[37]WEN Yun,MIRAULT J,GRAINGER J.Fast syntax in the brain:electrophysiological evidence from the rapid parallel visual presentation paradigm (RPVP)[J].Journal of Experimental Psychology:Learning,Memory,and Cognition,2021,47(1):99-112.
[38]朱夢澤,趙海英.敘事式可視化綜述[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2019,31(10):1719-1727.
ZHU Mengze,ZHAO Haiying.A summary of narrative visualization[J].Journal of Computer-Aided Design & Computer Graphics,2019,31(10):1719-1727.
[39]梁婷婷,邱素貞,陸珊.本科層次大數(shù)據(jù)人才需求可視化分析[J].智能計(jì)算機(jī)與應(yīng)用,2020,10(4):276-279.
LIANG Tingting,QIU Suzhen,LU Shan.Visual analysis of undergraduate level big data talent demand[J].Intelligent Computer and Applications,2020,10(4):276-279.
[40]陶芳.網(wǎng)絡(luò)輿論數(shù)據(jù)可視化技術(shù)研究[D].北京:北京郵電大學(xué),2019.
TAO Fang.Research on Network Public Opinion Data Visualization Technology[D].Beijing:Beijing University of Posts and Telecommunications,2019.
[41]張?zhí)鹛?,李妮,龔光紅,等.一種基于數(shù)獨(dú)分組的拉丁超立方試驗(yàn)設(shè)計(jì)方法[J].系統(tǒng)仿真學(xué)報(bào),2020,32(11):2185-2191.
ZHANG Tiantian,LI Ni,GONG Guanghong,et al.A fast latin hyper cube experiment design method based on Soduku grouping[J].Journal of System Simulation,2020,32(11):2185-2191.
[42]ZHANG Lidong,VINODHINI B,MARAGATHAM T.Interactive IOT data visualization for decision making in business intelligence[J].Arabian Journal for Science and Engineering,2021,70:1-11.
[43]崔卓,賈力,黃豆,等.矩形微槽道平板熱管流動與傳熱特性可視化[J].科學(xué)通報(bào),2020,65(17):1770-1779.
CUI Zhuo,JIA Li,HUANG Dou,et al.The visualized investigation on flow and heat transfer characteristics of flat-plate heat pipe with micro rectangular grooves[J].Chinese Science Bulletin,2020,65(17):1770-1779.
[44]孟鑫,李洋,王學(xué)良.多組件Web可視化開發(fā)平臺在調(diào)度自動化系統(tǒng)中的應(yīng)用[J].計(jì)算機(jī)系統(tǒng)應(yīng)用,2020,29(12):100-105.
MENG Xin,LI Yang,WANG Xueliang.Application of multi-component Web visual development platform in scheduling automation system[J].Computer Systems & Applications,2020,29(12):100-105.
[45]張慧軍,陳俊杰,張小龍.利用問題求解理論來研究交互式復(fù)雜信息的可視分析行為[J].圖學(xué)學(xué)報(bào),2020,41(3):325-334.
ZHANG Huijun,CHEN Junjie,ZHANG Xiaolong.Using problem-solving theories to investigate user behaviors in interactive visual analytics of complex information[J].Journal of Graphics,2020,41(3):325-334.
[46]KRAUS M,KLEIN K,F(xiàn)UCHS J,et al.The value of immersive visualization[J].IEEE Computer Graphics and Applications,2021,41(4):125-132.
[47]LIU Zhansheng,BAI Wenyan.Building information modeling methods for post-earthquake retrofitting visualization of buildings using augmented reality[J].Applied Sciences,2021,11(12):5739.
[48]崔蓬.ECharts在數(shù)據(jù)可視化中的應(yīng)用[J].軟件工程,2019,22(6):42-46.
CUI Peng.Study on the application of ECharts in data visualization[J].Software Engineering,2019,22(6):42-46.
[49]何巍.社交媒體數(shù)據(jù)可視化分析綜述[J].科學(xué)技術(shù)與工程,2020,20(32):13085-13090.
HE Wei.An overview on social media visual analytics[J].Science Technology and Engineering,2020,20(32):13085-13090.
[50]馬夢宇,吳燁,陳犖,等.顯示導(dǎo)向型的大規(guī)模地理矢量實(shí)時可視化技術(shù)[J].計(jì)算機(jī)科學(xué),2020,47(9):117-122.
MA Mengyu,WU Ye,CHEN Luo,et al.Display-oriented data visualization technique for large-scale geographic vector data[J].Computer Science,2020,47(9):117-122.
[51]沈恩亞.大數(shù)據(jù)可視化技術(shù)及應(yīng)用[J].科技導(dǎo)報(bào),2020,38(3):68-83.
SHEN Enya.Big data visualization technology and applications[J].Science & Technology Review,2020,38(3):68-83.
[52]LI Mengyao,DU Wenli,QIAN Feng,et al.Total plant performance evaluation based on big data:Visualization analysis of TE process[J].Chinese Journal of Chemical Engineering,2018,26(8):1736-1749.
[53]BAKHSHI A K,AHMED M M.Utilizing black-box visualization tools to interpret non-parametric real-time risk assessment models[J].Transportmetrica A:Transport Science,2020,17(4):739-765.
[54]邊根慶,陳蔚韜.面向Web的建筑三維模型可視化方法研究[J].圖學(xué)學(xué)報(bào),2021,42(5):823-832.
BIAN Genqing,CHEN Weitao.Research on web-oriented visualization method of 3D architectural model[J].Journal of Graphics,2021,42(5):823-832.
[55]駱菁菁,唐衛(wèi)貞,丁繼婷.基于皮爾遜系數(shù)的管制仿真訓(xùn)練數(shù)據(jù)獨(dú)立化與因子分析下的數(shù)據(jù)可視化研究[J].計(jì)算機(jī)科學(xué),2021,48(sup1):623-628.
LUO Jingjing,TANG Weizhen,DING Jiting.Research of ATC simulator training values independence based on pearson correlation coefficient and study of data visualization based on factor analysis[J].Computer Science,2021,48(sup1):623-628.
[56]戴國忠,陳為,洪文學(xué),等.信息可視化和可視分析:挑戰(zhàn)與機(jī)遇——北戴河信息可視化戰(zhàn)略研討會總結(jié)報(bào)告[J].中國科學(xué):信息科學(xué),2013,43(1):178-184.
DAI Guozhong,CHEN Wei,HONG Wenxue,et al.Information visualization and visual analysis:Challenges and opportunities - summary report of beidaihe information visualization strategy seminar[J].Science China:Information Science,2013,43(1):178-184.
[57]JIA Hekun,LIU Zhiyuan,YIN Bifeng,et al.Visualization research and optimization strategy for combustion process and emission characteristics of internal exhaust gas recirculation small non-road diesel engine[J].International Journal of Automotive Technology,2020,21(2):481-492.
[58]黃河燕,曹朝,馮沖.大數(shù)據(jù)情報(bào)分析發(fā)展機(jī)遇及其挑戰(zhàn)[J].智能系統(tǒng)學(xué)報(bào),2016,11(6):719-727.
HUANG Heyan,CAO Chao,F(xiàn)ENG Chong.Opportunities and challenges of big data intelligence analysis[J].Journal of Intelligent Systems,2016,11(6):719-727.
[59]張新榮,馬杰,張才斗,等.基于MATLAB倒立擺可視化建模仿真與控制[J].計(jì)算機(jī)工程與設(shè)計(jì),2018,39(10):3214-3219.
ZHANG Xinrong,MA Jie,ZHANG Caidou,et al.Visualized modeling and simulation of inverted pendulum control based on MATLAB[J].Computer Engineering and Design,2018,39(10):3214-3219.
[60]LI Mengyao,DU Wenli,QIAN Feng,et al.Total plant performance evaluation based on big data:Visualization analysis of TE process[J].Chinese Journal of Chemical Engineering,2018,26(8):1736-1749.
[61]孫燦燃,崔晨,徐靜樂,等.基于虛擬現(xiàn)實(shí)的社交網(wǎng)絡(luò)三維可視化系統(tǒng)[J].計(jì)算機(jī)仿真,2019,36(7):318-322.
SUN Canran,CUI Chen,XU Jingle,et al.Social network 3D visualization system based on virtual reality[J].Computer Simulation,2019,36(7):318-322.
[62]LEE B,ISAACS K,SZAFIR D A,et al.Broadening intellectual diversity in visualization research papers[J].IEEE Computer Graphics and Applications,2019,39(4):78-85.
收稿日期:2021-09-16;修回日期:2021-10-28;責(zé)任編輯:張士瑩
基金項(xiàng)目:國家文化和旅游科技創(chuàng)新工程項(xiàng)目(2020年度);河北省省級科技計(jì)劃資助項(xiàng)目(20310802D,21310101D,20310701D);河北省社會科學(xué)基金項(xiàng)目(HB20TQ008);河北省高層次人才資助項(xiàng)目(A2016002015);河北省創(chuàng)新能力提升計(jì)劃項(xiàng)目(20551801K);石家莊市科學(xué)技術(shù)研究與發(fā)展計(jì)劃項(xiàng)目(19SCX01006,191130591A)
第一作者簡介:劉 濱(1975—),男,河北石家莊人,教授,博士,主要從事大數(shù)據(jù)、社會計(jì)算方面的研究。
通訊作者:劉 宇副研究館員。E-mail:446927742@qq.com