楊彥波,劉 濱,祁明月
(1.河北科技大學(xué)經(jīng)濟(jì)管理學(xué)院,河北石家莊 050018;2.河北省軍區(qū)通信站,河北石家莊 050011)
信息可視化研究綜述
楊彥波1,劉 濱1,祁明月2
(1.河北科技大學(xué)經(jīng)濟(jì)管理學(xué)院,河北石家莊 050018;2.河北省軍區(qū)通信站,河北石家莊 050011)
信息可視化是可視化技術(shù)在非空間數(shù)據(jù)領(lǐng)域的應(yīng)用,可以增強(qiáng)數(shù)據(jù)呈現(xiàn)效果,讓用戶以直觀交互的方式實(shí)現(xiàn)對(duì)數(shù)據(jù)的觀察和瀏覽,從而發(fā)現(xiàn)數(shù)據(jù)中隱藏的特征、關(guān)系和模式??梢暬瘧?yīng)用非常廣泛,主要涉及領(lǐng)域:數(shù)據(jù)挖掘可視化、網(wǎng)絡(luò)數(shù)據(jù)可視化、社交可視化、交通可視化、文本可視化、生物醫(yī)藥可視化等等。根據(jù)CARD可視化模型可以將信息可視化的過(guò)程分為以下幾個(gè)階段:數(shù)據(jù)預(yù)處理;繪制;顯示和交互。根據(jù)SHNEIDERMAN的分類,信息可視化的數(shù)據(jù)分為以下幾類:一維數(shù)據(jù)、二維數(shù)據(jù)、三維數(shù)據(jù)、多維數(shù)據(jù)、時(shí)態(tài)數(shù)據(jù)、層次數(shù)據(jù)和網(wǎng)絡(luò)數(shù)據(jù)。其中針對(duì)后4種數(shù)據(jù)的可視化是當(dāng)前研究的熱點(diǎn)。
多維數(shù)據(jù)可視化方法主要包括基于幾何的方法、圖標(biāo)方法和動(dòng)畫(huà)方法等?;趲缀蔚目梢暬绞街凶罱?jīng)典的就是“平行坐標(biāo)系”方法。平行坐標(biāo)系(parallel coordinates)使用平行的豎直軸線來(lái)代表維度,通過(guò)在軸上刻劃多維數(shù)據(jù)的數(shù)值并用折線相連某一數(shù)據(jù)項(xiàng)在所有軸上的坐標(biāo)點(diǎn)展示多維數(shù)據(jù)。平行坐標(biāo)系方法能夠簡(jiǎn)潔、快速地展示多維數(shù)據(jù),發(fā)展出很多改進(jìn)技術(shù)。但是當(dāng)數(shù)據(jù)集的規(guī)模變得非常大時(shí),密集的折線會(huì)引起“視覺(jué)混淆”(visual clutter),處理方法包括維度重排、交互方法、聚類、過(guò)濾、動(dòng)畫(huà)等。其他基于幾何的方法包括Radviz方法使用圓形坐標(biāo)系展示可視化結(jié)果;散點(diǎn)圖矩陣(scatter plot matrix)將多維數(shù)據(jù)中的各個(gè)維度兩兩組合繪制成一系列的按規(guī)律排列的散點(diǎn)圖?;趫D標(biāo)的可視化方法用具備可視特征的幾何形狀如大小、長(zhǎng)度、形狀、顏色等刻劃數(shù)據(jù),代表性的方法包括星繪法和Chernoff 面法等。動(dòng)畫(huà)方法用于可視化中可被用來(lái)提高交互性和理解程度,其缺點(diǎn)包括可能分散注意力、引起用戶的誤解、產(chǎn)生“圖表垃圾”等。
時(shí)間序列數(shù)據(jù)是指具有時(shí)間屬性的數(shù)據(jù)集,針對(duì)時(shí)間序列數(shù)據(jù)的可視化方法如下:線形圖、堆積圖、動(dòng)畫(huà)、地平線圖、時(shí)間線。
層次數(shù)據(jù)具有等級(jí)或?qū)蛹?jí)關(guān)系。層次數(shù)據(jù)的可視化方法主要包括節(jié)點(diǎn)鏈接圖和樹(shù)圖2種方式。其中樹(shù)圖(treemap)由一系列的嵌套環(huán)、塊來(lái)展示層次數(shù)據(jù)。為了能展示更多的節(jié)點(diǎn)內(nèi)容,一些基于“焦點(diǎn)+上下文”技術(shù)的交互方法被開(kāi)發(fā)出來(lái)。包括“魚(yú)眼”技術(shù)、幾何變形、語(yǔ)義縮放、遠(yuǎn)離焦點(diǎn)的節(jié)點(diǎn)聚類技術(shù)等。
網(wǎng)絡(luò)數(shù)據(jù)具有網(wǎng)狀結(jié)構(gòu)。自動(dòng)布局算法是網(wǎng)絡(luò)數(shù)據(jù)可視化的核心,目前主要有以下3類:一是力導(dǎo)向布局(force-directed layout);二是分層布局(hierarchical layout);三是網(wǎng)格布局(grid layout)。當(dāng)數(shù)據(jù)節(jié)點(diǎn)的連接很多時(shí),容易產(chǎn)生邊交叉現(xiàn)象,導(dǎo)致視覺(jué)混淆。解決邊交叉現(xiàn)象的集束邊(edge bundle)技術(shù)可以分為以下幾類:力導(dǎo)向的集束邊技術(shù)、層次集束邊技術(shù)、基于幾何的邊聚類技術(shù)、多層凝聚集束邊技術(shù)和基于網(wǎng)格的方法等。
其他研究熱點(diǎn)包括圖形的視覺(jué)因素研究、自適應(yīng)可視化研究、可視化效果的評(píng)估等。
視覺(jué)因素對(duì)于可視化效果的影響,如位置、長(zhǎng)度、面積、形狀、色彩等影響已經(jīng)引起很多研究者的注意。色彩是視覺(jué)因素的重要組成部分,研究主要集中在顏色選擇的原則和交互系統(tǒng)中。這些原則基于數(shù)據(jù)類型、類的數(shù)量、認(rèn)知約束等。
自適應(yīng)可視化可以提高信息可視化的適應(yīng)性。研究成果分為以下幾類:自適應(yīng)可視化展示、自適應(yīng)資源模型、自適應(yīng)用戶模型。自適應(yīng)可視化展示是指根據(jù)用戶的特征自動(dòng)為用戶提供多種展示類型,自動(dòng)選擇可視化內(nèi)容及布局的形式,自動(dòng)調(diào)整可視化的元素等。自適應(yīng)資源模型反映了對(duì)硬件和軟件的利用以提高可視化性能。自適應(yīng)用戶模型通過(guò)顯示用戶模型的內(nèi)容并讓用戶能夠編輯,從而讓用戶能夠控制模型的內(nèi)容。
當(dāng)前關(guān)于信息可視化評(píng)價(jià)的研究較少,少量研究也沒(méi)有提出直接和通用的可視化的評(píng)估方式,需要對(duì)信息可視化評(píng)價(jià)的理論基礎(chǔ)、方法和應(yīng)用做深入的研究。
可視化技術(shù)與應(yīng)用還應(yīng)該繼續(xù)向以下4個(gè)方面努力:直觀化、關(guān)聯(lián)化、藝術(shù)化、交互化。信息可視化技術(shù)的發(fā)展方向是協(xié)同(collaboration)、分析過(guò)程(analytics)、計(jì)算(computational)和意會(huì)(sense-making)。未來(lái)研究方向可以包括以下幾個(gè)內(nèi)容。
信息可視化和數(shù)據(jù)挖掘的緊密結(jié)合。為提高處理海量數(shù)據(jù)時(shí)的速度和效率和解決視覺(jué)混淆現(xiàn)象;必須運(yùn)用數(shù)據(jù)挖掘的公式和算法,對(duì)數(shù)據(jù)分析的過(guò)程及結(jié)果進(jìn)行可視化展現(xiàn)。
協(xié)同可視化。協(xié)同可視化領(lǐng)域的研究方向可以包括可視化接口設(shè)計(jì)、基于Web的可視化協(xié)同平臺(tái)開(kāi)發(fā)、協(xié)同可視化工作的視圖設(shè)計(jì)、協(xié)同可視化中的工作流管理及協(xié)同可視化技術(shù)的應(yīng)用等。
更多領(lǐng)域的應(yīng)用技術(shù)開(kāi)發(fā)。包括統(tǒng)計(jì)可視化:需要研究使用幾何、動(dòng)畫(huà)、圖像等工具對(duì)數(shù)據(jù)統(tǒng)計(jì)的過(guò)程和結(jié)果進(jìn)行加工和處理的技術(shù);新聞可視化:對(duì)新聞內(nèi)容進(jìn)行抓取、清洗和提取和可視化展示;社交網(wǎng)絡(luò)可視化:可視化方式顯示社交網(wǎng)絡(luò)的數(shù)據(jù),對(duì)社交網(wǎng)絡(luò)中節(jié)點(diǎn)、關(guān)系及時(shí)空數(shù)據(jù)的集成展示。搜索日志可視化:針對(duì)在使用搜索引擎時(shí)產(chǎn)生的海量搜索日志,可視化的展現(xiàn)用戶的搜索行為、關(guān)系和模式等。
信息可視化;可視化技術(shù);人機(jī)交互;數(shù)據(jù)挖掘
可視化技術(shù)起源于20世紀(jì)80年代出現(xiàn)的科學(xué)計(jì)算可視化[1]?!靶畔⒖梢暬币辉~最早出現(xiàn)在ROBERTSON等1989 年發(fā)表的文章《用于交互性用戶界面的認(rèn)知協(xié)處理器》中[2]。信息可視化是可視化技術(shù)在非空間數(shù)據(jù)領(lǐng)域的應(yīng)用,是將數(shù)據(jù)信息轉(zhuǎn)化為視覺(jué)形式的過(guò)程,可以增強(qiáng)數(shù)據(jù)呈現(xiàn)效果,讓用戶以直觀交互的方式實(shí)現(xiàn)對(duì)數(shù)據(jù)的觀察和瀏覽,從而發(fā)現(xiàn)數(shù)據(jù)中隱藏的特征、關(guān)系和模式。信息可視化的圖表形式最早出現(xiàn)于18世紀(jì),歷史和政治學(xué)家PLAYFAIR和數(shù)學(xué)家LAMBERT首次創(chuàng)建了可視化圖表,他們認(rèn)為將復(fù)雜的數(shù)據(jù)轉(zhuǎn)化為圖表可以幫助人們了解數(shù)據(jù)。19世紀(jì)的法國(guó)科學(xué)家MINARD和MAREY首次采用非純手工方式繪制了圖表[3]。進(jìn)入20世紀(jì),現(xiàn)在計(jì)算機(jī)技術(shù)的進(jìn)步拓展了數(shù)據(jù)處理的能力并且可以提供多種交互方式,使得用戶可以更便利的觀察自己感興趣的數(shù)據(jù),可視化應(yīng)用也更加廣泛,主要領(lǐng)域涉及數(shù)據(jù)挖掘可視化、網(wǎng)絡(luò)數(shù)據(jù)可視化、社交可視化、交通可視化、文本可視化、生物醫(yī)藥可視化等。
1.1Card信息可視化模型
在CARD等提出的信息可視化模型中(見(jiàn)圖1),信息可視化過(guò)程可以劃分為3個(gè)數(shù)據(jù)轉(zhuǎn)換的過(guò)程:原始數(shù)據(jù)到數(shù)據(jù)表的轉(zhuǎn)換、數(shù)據(jù)表到可視化結(jié)構(gòu)的轉(zhuǎn)換、可視化結(jié)構(gòu)到視圖的轉(zhuǎn)換[4]。
圖1 Card信息可視化模型Fig.1 Card model of information visualization
1.2信息可視化過(guò)程
根據(jù)Card信息可視化模型可以將信息可視化的過(guò)程分為以下3個(gè)階段[5]。
1)數(shù)據(jù)預(yù)處理 數(shù)據(jù)預(yù)處理將采集來(lái)的信息進(jìn)行預(yù)處理和加工,使其便于理解,易于被輸入顯示可視化模塊。預(yù)處理內(nèi)容包括數(shù)據(jù)格式及其標(biāo)準(zhǔn)化、數(shù)據(jù)變換技術(shù)、數(shù)據(jù)壓縮和解壓縮等。有些數(shù)據(jù)也需要做異常值檢出、聚類、降維等處理。
2)繪制 繪制的功能是完成數(shù)據(jù)到幾何圖像的轉(zhuǎn)換。一個(gè)完整的圖形描述需要在考慮用戶需求的基礎(chǔ)上綜合應(yīng)用各類可視化繪制技術(shù)。
3)顯示和交互 顯示的功能是將繪制模塊生成的圖像數(shù)據(jù),按用戶指定的要求進(jìn)行輸出。除了完成圖像信息輸出功能外,還需要把用戶的反饋信息傳送到軟件層中,以實(shí)現(xiàn)人機(jī)交互。針對(duì)可視化的主要任務(wù),即總覽(overview)、縮放(zoom)、過(guò)濾(filter)、詳細(xì)查看(details-on-demand)、關(guān)聯(lián)(relate)等,交互技術(shù)主要包括動(dòng)態(tài)過(guò)濾、全局+詳細(xì)、平移+縮放、焦點(diǎn)+上下文及變形、多視圖關(guān)聯(lián)協(xié)調(diào)等技術(shù)[6-7]。
根據(jù)SHNEIDERMAN的分類,信息可視化的數(shù)據(jù)分為以下幾類:一維數(shù)據(jù);二維數(shù)據(jù);三維數(shù)據(jù);多維數(shù)據(jù);時(shí)態(tài)數(shù)據(jù);層次數(shù)據(jù)和網(wǎng)絡(luò)數(shù)據(jù)[8]。其中針對(duì)后4種數(shù)據(jù)的可視化是當(dāng)前研究的熱點(diǎn),本文將就這4種數(shù)據(jù)的可視化方法展開(kāi)綜述。
2.1多維數(shù)據(jù)可視化技術(shù)
針對(duì)多維數(shù)據(jù),采用傳統(tǒng)二維圖表方式難以有效滿足現(xiàn)代化的大量、復(fù)雜、多維度的信息需求。多維數(shù)據(jù)的可視化是當(dāng)前研究的熱點(diǎn)之一。多維數(shù)據(jù)可視化的方法有很多種,本文主要綜述和討論有代表性的幾種方法。
圖2 平行坐標(biāo)系Fig.2 Parallel coordinates
2.1.1 基于幾何的可視化方式
1)平行坐標(biāo)系 1980年,INSELBERG提出的平行坐標(biāo)系(parallel coordinates)是經(jīng)典的多維數(shù)據(jù)可視化技術(shù)之一(見(jiàn)圖2)[9]。平行坐標(biāo)系使用平行的豎直軸線來(lái)代表維度,通過(guò)在軸上刻劃多維數(shù)據(jù)的數(shù)值并用折線相連某一數(shù)據(jù)項(xiàng)在所有軸上的坐標(biāo)點(diǎn),從而在二維空間內(nèi)展示多維數(shù)據(jù)。
平行坐標(biāo)系方法能夠簡(jiǎn)潔、快速地展示多維數(shù)據(jù)。由于其經(jīng)典性和廣泛應(yīng)用性,許多學(xué)者將平行坐標(biāo)系法應(yīng)用于可視化、數(shù)據(jù)挖掘、過(guò)程控制、決策支持、近似計(jì)算和其他一些領(lǐng)域并獲得成功[10]。1990年INSELBERG首先將平行坐標(biāo)系法用來(lái)解決可視化問(wèn)題后[11],平行坐標(biāo)系法也發(fā)展出很多改進(jìn)技術(shù)[12],比如在不同層次上的平行坐標(biāo)顯示[13],用曲線代替直線增強(qiáng)可視化效果等[14]。盛秀杰等使用平行坐標(biāo)中的坐標(biāo)軸和平行折線的可視化渲染方法提出了一種新的顏色漸變渲染方案[15]。SIIRTOLA提出利用數(shù)據(jù)子集的相關(guān)系數(shù)的平均數(shù)的方法動(dòng)態(tài)畫(huà)出折線[16]。WONG等使用小波逼近方法建立的涂刷工具能夠展示不同分辨率下的線條構(gòu)成[17]。Angular brushing可以讓用戶方便地選擇2個(gè)數(shù)軸之間的數(shù)據(jù)子集[18]。EdgeLens能夠在保證節(jié)點(diǎn)完好不變形的情況下交互地展示焦點(diǎn)區(qū)域中心的折線[19]。ZHOU等使用可視化聚類的方法調(diào)整折線的形狀[20]。
平行坐標(biāo)可以進(jìn)一步擴(kuò)展到三維可視化的方式以展示高維動(dòng)態(tài)的數(shù)據(jù)。很多專家也把平行坐標(biāo)系和其他方法結(jié)合。SpringView整合了平行坐標(biāo)系法和放射坐標(biāo)系法來(lái)解決多維數(shù)據(jù)集。Parallel Glyphs將各個(gè)坐標(biāo)軸擴(kuò)展到星形圖的空間中以方便進(jìn)行數(shù)據(jù)對(duì)比和提供交互[21]。
當(dāng)數(shù)據(jù)集的規(guī)模變得非常大時(shí),密集的折線會(huì)讓平行坐標(biāo)系變得難以解釋。因此降低視覺(jué)混淆也被很多專家關(guān)注,基于平行坐標(biāo)系的視覺(jué)混淆處理方法包括維度重排[22]、交互方法[20]、聚類[23]、過(guò)濾[24]、動(dòng)畫(huà)[25]等。
圖3 Radviz方法Fig.3 Radviz method
圖4 散點(diǎn)圖矩陣Fig.4 Scatter plot matrix
2)Radviz方法 Radviz(radial coordinate visualization)使用圓形坐標(biāo)系展示可視化結(jié)果(見(jiàn)圖3),圓形的k條半徑表示k維空間,通過(guò)引入物理學(xué)中物體受力平衡定理,將多維數(shù)據(jù)對(duì)象表示為坐標(biāo)系內(nèi)的一個(gè)點(diǎn),點(diǎn)的位置使用彈簧模型計(jì)算得到[26]。Radviz的優(yōu)缺點(diǎn)同平行坐標(biāo)系類似,當(dāng)數(shù)據(jù)規(guī)模很大時(shí),也容易產(chǎn)生視覺(jué)混淆現(xiàn)象,影響用戶對(duì)于可視化結(jié)果的認(rèn)知。
3)散點(diǎn)圖矩陣 散點(diǎn)圖通過(guò)二維坐標(biāo)系中的一組點(diǎn)來(lái)展示2個(gè)變量之間的關(guān)系,散點(diǎn)圖矩陣(scatter plot matrix)就是將多維數(shù)據(jù)中的各個(gè)維度兩兩組合繪制成一系列的按規(guī)律排列的散點(diǎn)圖(見(jiàn)圖4)。散點(diǎn)圖矩陣也經(jīng)常和其他可視化方法結(jié)合來(lái)增強(qiáng)顯示多維數(shù)據(jù)效果,基于散點(diǎn)圖矩陣的開(kāi)發(fā)的連續(xù)的散點(diǎn)圖可以對(duì)海量數(shù)據(jù)進(jìn)行可視化展示,CRAIG等研究了傳統(tǒng)的時(shí)間序列圖和散點(diǎn)圖的互補(bǔ)關(guān)系[27],SCHMID等整合了散點(diǎn)圖矩陣、平行坐標(biāo)系、Addrews曲線來(lái)展示多維數(shù)據(jù)[28]。散點(diǎn)圖矩陣的優(yōu)點(diǎn)主要是能快速發(fā)現(xiàn)成對(duì)變量之間的關(guān)系,缺點(diǎn)是當(dāng)數(shù)據(jù)維度太大時(shí),屏幕的大小會(huì)限制顯示矩陣元素的數(shù)量,需要結(jié)合交互技術(shù)來(lái)實(shí)現(xiàn)用戶對(duì)可視化結(jié)果的觀察。
圖5 Andrews曲線Fig.5 Andrews curve
4)Andrews 曲線法 Andrews曲線法(見(jiàn)圖5)使用二維坐標(biāo)系展示可視化結(jié)果,將多維數(shù)據(jù)的每一數(shù)據(jù)項(xiàng)通過(guò)一個(gè)周期函數(shù)映射到二維坐標(biāo)系中的一條曲線上[29],通過(guò)對(duì)曲線的觀察,用戶能夠感知數(shù)據(jù)的聚類等狀況。
2.1.2 基于圖標(biāo)的可視化方式
基于圖標(biāo)的可視化方法用具備可視特征的幾何形狀作為圖標(biāo)來(lái)刻劃多維數(shù)據(jù),這些圖標(biāo)的每一個(gè)可視化屬性如大小、長(zhǎng)度、形狀、顏色可以作為維度,通過(guò)多維數(shù)據(jù)到這些圖標(biāo)屬性的映射來(lái)實(shí)現(xiàn)可視化效果,代表性的方法包括星繪法(見(jiàn)圖6)和Chernoff 面法(見(jiàn)圖7)等。星繪法采用由一點(diǎn)向外輻射的多條線段代表數(shù)據(jù)維度,不同的線段長(zhǎng)度代表了每一個(gè)數(shù)據(jù)項(xiàng)的不同維度的值。Chernoff 面法使用人臉的大小、形狀和臉部器官的特征來(lái)代表數(shù)據(jù)維度,通過(guò)人臉繪制的多維數(shù)據(jù)按一定的策略進(jìn)行排序,可以實(shí)現(xiàn)數(shù)據(jù)的可視化展示[30]。Chernoff 面法因?yàn)樵谡宫F(xiàn)上比其他圖形技術(shù)更有趣,所以觀察者愿意花更多的時(shí)間去分析,因此展現(xiàn)可能會(huì)更有效,Chernoff 面法也有利于高效識(shí)別各個(gè)要素之間的關(guān)系或模式[31]。
圖6 星繪法Fig.6 Star graph
圖7 Chernoff 面法Fig.7 Chernoff face
2.1.3 動(dòng)畫(huà)的多維可視化技術(shù)研究現(xiàn)狀
由于具有直觀和引人入勝的特點(diǎn),動(dòng)畫(huà)已經(jīng)被廣泛應(yīng)用于用戶界面中。文獻(xiàn)表明,動(dòng)畫(huà)可被用來(lái)提高交互性和理解程度:1)運(yùn)動(dòng)的物體能夠有效的吸引人們的注意力;2)動(dòng)畫(huà)能展現(xiàn)對(duì)象的漸變[32],包括位置、大小、形狀、顏色的變化,從而讓人能自然地感覺(jué)到對(duì)象的變化;3)動(dòng)畫(huà)可以提高用戶對(duì)因果關(guān)系和指向性的感知;4)動(dòng)畫(huà)可以提升用戶的興趣度,讓用戶能更享受瀏覽過(guò)程[33]。
如何設(shè)計(jì)動(dòng)畫(huà)以方便用戶的理解是研究的熱點(diǎn)之一。一種方法是在對(duì)數(shù)據(jù)進(jìn)行編碼時(shí)使用動(dòng)作作為一個(gè)附加的視覺(jué)變量。另一種方法是利用動(dòng)畫(huà)使不同狀態(tài)下的轉(zhuǎn)換容易理解,如樹(shù)圖可視化中動(dòng)畫(huà)轉(zhuǎn)換的應(yīng)用。錐樹(shù)在樹(shù)的多個(gè)層次上利用動(dòng)畫(huà)的旋轉(zhuǎn)使選定的項(xiàng)目進(jìn)入觀察者的視野[34]。SpaceTrees和DOITrees利用動(dòng)畫(huà)動(dòng)態(tài)地展示樹(shù)的枝葉展開(kāi)和折疊的情形。動(dòng)畫(huà)也被應(yīng)用于統(tǒng)計(jì)圖表中[35-36]。The Name Voyager基于Many Eyes技術(shù)在堆疊面積圖中使用動(dòng)畫(huà)的方式展現(xiàn)尺度、網(wǎng)格線和軸標(biāo)簽的變化[37]。BUSH等使用動(dòng)畫(huà)介紹和分析統(tǒng)計(jì)圖,包括數(shù)據(jù)標(biāo)記的動(dòng)態(tài)展現(xiàn)、從堆積面積圖到散點(diǎn)圖變形和轉(zhuǎn)換等[38]。
有一些專家也指出了動(dòng)畫(huà)的缺點(diǎn)。如動(dòng)畫(huà)可能分散注意力、動(dòng)畫(huà)的在對(duì)象轉(zhuǎn)換的過(guò)程中可能會(huì)引起用戶的誤解,可能用戶注意“圖表垃圾”(不相關(guān)的信息和無(wú)價(jià)值的信息)[39]。另外,如果動(dòng)畫(huà)速度太慢會(huì)浪費(fèi)時(shí)間,動(dòng)畫(huà)速度太快會(huì)導(dǎo)致理解錯(cuò)誤,但動(dòng)畫(huà)的最佳速度由于場(chǎng)景復(fù)雜性和用戶的背景很難確定。因此,對(duì)于動(dòng)畫(huà)的使用必須慎重。
2.2時(shí)間序列數(shù)據(jù)的可視化
時(shí)間序列數(shù)據(jù)是指具有時(shí)間屬性的數(shù)據(jù)集,針對(duì)時(shí)間序列數(shù)據(jù)的可視化方法如下。
1)線形圖 線形圖是時(shí)間序列可視化中最普通的方式,使用點(diǎn)的位置代表時(shí)間發(fā)展和數(shù)據(jù)值。對(duì)于有多個(gè)時(shí)間維度的數(shù)據(jù)可以為每一個(gè)時(shí)間維建立一個(gè)圖表,可以讓圖表垂直和水平對(duì)齊,以幫助進(jìn)行事件的趨勢(shì)比較。
2)堆積圖 堆積圖是對(duì)時(shí)間序列數(shù)據(jù)累積形式的展現(xiàn),可以觀察序列的總和。堆積圖雖然能夠有效地顯示序列總和模式,但是缺乏每個(gè)序列的比較,處理含有負(fù)值的數(shù)據(jù)也較差。
圖8 地平線圖Fig.8 Horizon graph
3)動(dòng)畫(huà) 和靜態(tài)圖片相比,動(dòng)畫(huà)方式更能展現(xiàn)時(shí)間序列的變化情況。但是動(dòng)畫(huà)精確顯示數(shù)值的能力較差。
4)地平線圖 地平線圖能夠展現(xiàn)數(shù)據(jù)的變化率隨時(shí)間的演變情況,并且可以使用顏色來(lái)加深正向變動(dòng)和負(fù)向變動(dòng)的效果(見(jiàn)圖8)。SAITO等首次提出地平線圖技術(shù)[40],HEER等進(jìn)一步發(fā)展了這項(xiàng)技術(shù)[41]。
圖9 iPhone發(fā)展時(shí)間線Fig.9 Timeline of iPhone3
5)時(shí)間線 時(shí)間線(Timeline)是指以時(shí)間軸為水平軸線,將數(shù)據(jù)信息以圖標(biāo)或圖片的形式按時(shí)間順序置于水平軸坐標(biāo)系內(nèi)。1765年,PRIESTLEY用時(shí)間線的方法描述了從公元前1200年到公元1750年間的2 000位著名人士的生命期內(nèi)的事件。時(shí)間線也被用在了醫(yī)療記錄和犯罪記錄中,最著名的是Lifelines。Lifelines 使用時(shí)間線展示人的歷史醫(yī)療情況,并且可以點(diǎn)擊事件點(diǎn)查看詳細(xì)的信息。The Pattern Finder是Lifelines的發(fā)展,用于多維數(shù)據(jù)的可視化挖掘中[42]。時(shí)間線的最主要的問(wèn)題就是由于時(shí)間范圍過(guò)長(zhǎng)從而難以在長(zhǎng)度有限的時(shí)間軸上全面展示重要的信息細(xì)節(jié)。為了解決這個(gè)問(wèn)題,SPENCE基于“焦點(diǎn)+上下文”技術(shù)提出的“Bifocal Lens”方法可以從語(yǔ)義上壓縮顯示邊緣的顯示項(xiàng)目[43]。MACKINLAY等提出了另一種基于“焦點(diǎn)+上下文”技術(shù)的方法“Perspective Wall”使用透視方法壓縮顯示邊緣的信息[44]。將不同時(shí)間線聯(lián)系起來(lái)也能展現(xiàn)更復(fù)雜的關(guān)系。JENSEN通過(guò)將多維時(shí)間軸能夠堆疊和鏈接的方式來(lái)展現(xiàn)不同時(shí)間線的事件之間的關(guān)系[45]。如iPhone發(fā)展時(shí)間線見(jiàn)圖9。
2.3層次數(shù)據(jù)的可視化
層次數(shù)據(jù)是常見(jiàn)的數(shù)據(jù)類型,可以用來(lái)描述生命物種、組織結(jié)構(gòu)、家庭關(guān)系、社會(huì)網(wǎng)絡(luò)等具有等級(jí)或?qū)蛹?jí)關(guān)系的對(duì)象。層次數(shù)據(jù)的可視化方法主要包括節(jié)點(diǎn)鏈接圖和樹(shù)圖2種方式。
圖10 節(jié)點(diǎn)鏈接圖Fig.10 Link point graph
圖11 樹(shù)圖Fig.11 Treemap4
1)節(jié)點(diǎn)鏈接圖 節(jié)點(diǎn)鏈接圖[46]是將層次數(shù)據(jù)組織成一個(gè)類似于樹(shù)的節(jié)點(diǎn)的連接結(jié)構(gòu),畫(huà)出節(jié)點(diǎn)和連線來(lái)代表數(shù)據(jù)項(xiàng)和它們之間的關(guān)系,節(jié)點(diǎn)通常是一些小點(diǎn)從而難以包含更多的信息(見(jiàn)圖10)。節(jié)點(diǎn)鏈接圖能清晰直觀地展現(xiàn)層次數(shù)據(jù)內(nèi)的關(guān)系,但是分支間的空白會(huì)浪費(fèi)展示空間,當(dāng)數(shù)據(jù)量較大時(shí),分支很快就會(huì)擁擠交織在一起,變得混亂不堪,造成視覺(jué)混淆現(xiàn)象。
2)樹(shù)圖 樹(shù)圖(treemap)最早由JOHNSON等在1991年提出[47]。樹(shù)圖一系列的嵌套環(huán)、塊來(lái)展示層次數(shù)據(jù),樹(shù)圖能夠在有限的空間內(nèi)展示大量數(shù)據(jù),但是也無(wú)法展示節(jié)點(diǎn)的細(xì)節(jié)內(nèi)容(見(jiàn)圖11)。為了能展示更多的節(jié)點(diǎn)內(nèi)容,一些基于“焦點(diǎn)+上下文”技術(shù)的交互方法被開(kāi)發(fā)出來(lái)。包括“魚(yú)眼”技術(shù)、幾何變形、語(yǔ)義縮放、遠(yuǎn)離焦點(diǎn)的節(jié)點(diǎn)聚類技術(shù)等[48]。
2.4網(wǎng)絡(luò)數(shù)據(jù)可視化
網(wǎng)絡(luò)數(shù)據(jù)具有網(wǎng)狀結(jié)構(gòu),如互聯(lián)網(wǎng)網(wǎng)絡(luò)、社交網(wǎng)絡(luò)、合作網(wǎng)絡(luò)及傳播網(wǎng)絡(luò)等。自動(dòng)布局算法是網(wǎng)絡(luò)數(shù)據(jù)可視化的核心,目前主要有以下3類:一是仿真物理學(xué)中力的概念來(lái)繪制網(wǎng)狀圖,即力導(dǎo)向布局(force-directed layout);二是分層布局(hierarchical layout);三是網(wǎng)格布局(grid layout)。很多研究是基于以上布局算法的應(yīng)用或者是對(duì)以上算法的進(jìn)一步優(yōu)化。在網(wǎng)絡(luò)數(shù)據(jù)的可視化中,當(dāng)數(shù)據(jù)節(jié)點(diǎn)的連接很多時(shí),容易產(chǎn)生邊交叉現(xiàn)象,導(dǎo)致視覺(jué)混淆。解決邊交叉現(xiàn)象的集束邊(edge bundle)技術(shù)可以分為以下幾類:力導(dǎo)向的集束邊技術(shù)、層次集束邊技術(shù)、基于幾何的邊聚類技術(shù)、多層凝聚集束邊技術(shù)和基于網(wǎng)格的方法等[49]。
3.1可視化圖形的視覺(jué)因素研究
視覺(jué)因素對(duì)于可視化效果的影響,如位置、長(zhǎng)度、面積、形狀、色彩等影響已經(jīng)引起很多研究者的注意。評(píng)估視覺(jué)因素對(duì)于用戶感知的影響,使設(shè)計(jì)人員能夠優(yōu)化可視化效果,是可視化過(guò)程中的重點(diǎn)之一。BERTIN首次進(jìn)行了系統(tǒng)的實(shí)驗(yàn)[50],研究了視覺(jué)要素在名義、序列和度量數(shù)據(jù)可視化中的效果,CLEVELAND等采用基于人類的主觀認(rèn)知的科學(xué)實(shí)驗(yàn)測(cè)試了可視化要素的影響[51]。SIMKIN等測(cè)試了人類對(duì)條形圖和餅形圖的認(rèn)知差異[52]。其他研究包括圖表的縱橫比研究[53]、數(shù)據(jù)軸的設(shè)計(jì)[54]、色彩選擇等。
色彩是視覺(jué)因素的重要組成部分,研究成果也較豐富。研究主要集中在顏色選擇的原則和交互系統(tǒng)中。這些原則基于數(shù)據(jù)類型、類的數(shù)量、認(rèn)知約束等。PRAVDAColor和ColorBrewer都根據(jù)數(shù)據(jù)類型來(lái)配置可視化效果中的色彩[55-56]。MEIER等使用藝術(shù)色彩理論來(lái)設(shè)計(jì)交互色彩的選擇[57]。COHEN-Or等使用同樣理論在圖片組合中提升色彩的和諧度[58]。色彩的交互研究主要集中在顏色繪圖基于空間頻率[49],認(rèn)知清晰度,色彩協(xié)調(diào)[59],和顯示能量消耗[60]。
其他學(xué)者也從不同方面研究了色彩的應(yīng)用,研究了如何為單詞和短語(yǔ)配置色彩。一些學(xué)者采用統(tǒng)計(jì)模型(如LDA,latent dirichlet allocation模型等)研究了人類對(duì)色彩名字的判斷模式,設(shè)計(jì)的工具允許用戶操縱色彩映射以直觀地搜索和過(guò)濾數(shù)據(jù)[61]。
3.2自適應(yīng)可視化研究
圖12 自適應(yīng)可視化模型Fig.12 Model of adaptive visualization
自適應(yīng)可視化可以提高信息可視化的適應(yīng)性,如依據(jù)對(duì)用戶行為分析后得出的用戶特征來(lái)自動(dòng)調(diào)整可視化形式,從而提高可視化結(jié)果的針對(duì)性。通過(guò)自適應(yīng)可視化技術(shù),能夠?yàn)橛脩粽{(diào)整可視化的要素和方式[62]。一些學(xué)者的研究成果分為以下幾類:自適應(yīng)可視化展示、自適應(yīng)資源模型、自適應(yīng)用戶模型(見(jiàn)圖12)[63-64]。
1)自適應(yīng)可視化展示 自適應(yīng)可視化展示是指根據(jù)用戶的特征自動(dòng)為用戶提供多種展示類型,自動(dòng)選擇進(jìn)行可視化內(nèi)容及布局的形式,自動(dòng)調(diào)整可視化的元素如顏色和圖標(biāo)等。ERST(external representation selection tutor)根據(jù)用戶背景知識(shí)和任務(wù)類型提供信息展示方式的選擇集[65];Opinion Space讓用戶更容易在高維空間內(nèi)看到他們的觀點(diǎn)[66];GANSNER等使用FDP(force directed placement)提出基于顏色配置變化的自適應(yīng)方法[67],KnowledgeSea使用SOM算法通過(guò)改變前景色和背景色、圖標(biāo)等元素來(lái)可視化展現(xiàn)教育的內(nèi)容[68]。
2)自適應(yīng)資源模型 自適應(yīng)資源模型反映了對(duì)硬件和軟件的利用以提高可視化性能。MEYER等開(kāi)發(fā)了一種能夠自動(dòng)利用網(wǎng)絡(luò)中多個(gè)服務(wù)器的計(jì)算機(jī)資源的實(shí)時(shí)優(yōu)化控制技術(shù)來(lái)提高交互可視化的性能[69]。GALLOP等提出用于支持分布環(huán)境下協(xié)同工作的面向組件的自適應(yīng)協(xié)同可視化技術(shù)[70]。LIU等研究了計(jì)算機(jī)負(fù)荷情況可視化過(guò)程的影響[71]。
3)自適應(yīng)用戶模型 自適應(yīng)用戶模型通過(guò)顯示用戶模型的內(nèi)容并讓用戶能夠編輯,從而讓用戶能夠控制模型的內(nèi)容。BRUSILOVSKY提出了如何建立用戶模型[72]。WOJCJECH等開(kāi)發(fā)了根據(jù)用戶搜索結(jié)果的屬性及用戶的相關(guān)信息來(lái)自適應(yīng)地選擇相應(yīng)界面的方法[73]。YourNews提供用戶模型的觀察和編輯功能。IntrospectiveView使用本體模型,根據(jù)用戶的不同提供不同顏色和字體[74]。MyExperiences引入開(kāi)放學(xué)習(xí)模型(OLM),從而允許用戶建構(gòu)自己的模型[75]。
上述自適應(yīng)可視化策略仍然多是靜態(tài)的,缺乏交互性,難以幫助用戶學(xué)習(xí)并理解復(fù)雜的信息需求[76]。
3.3可視化效果的評(píng)估
MORSE等指出當(dāng)前關(guān)于信息可視化評(píng)價(jià)的研究較少,少量研究也沒(méi)有提出直接和通用的可視化的評(píng)估方式[77]。MEYER指出沒(méi)有普遍接受的關(guān)于最優(yōu)數(shù)據(jù)展示的標(biāo)準(zhǔn),部分原因在于缺乏實(shí)證證據(jù)和可視化方法種類太多[78]。PURCHASE認(rèn)為評(píng)估只是根據(jù)美學(xué)因素或計(jì)算效率評(píng)價(jià)可視化是不可取的,而是應(yīng)該考慮可視化對(duì)用戶績(jī)效的提升能力。研究表明界面的美學(xué)特質(zhì)和績(jī)效之間的關(guān)系是復(fù)雜的[79]。雖然美學(xué)因素經(jīng)常和感知易用性相聯(lián)系,但是感知易用性和實(shí)際的可用性可能是不相關(guān)的。另一學(xué)者主張使用表現(xiàn)力(expressiveness) 和有效性(effectiveness) 兩個(gè)標(biāo)準(zhǔn)來(lái)評(píng)價(jià)信息可視化的效果。如FREITAS提出的認(rèn)知復(fù)雜度(cognitive complexity) 和空間組織(spatial organization) 2個(gè)標(biāo)準(zhǔn)[80]。
一些學(xué)者專注于特定領(lǐng)域的可視化的評(píng)估。DOWELL等認(rèn)為這些評(píng)價(jià)是針對(duì)特殊領(lǐng)域的評(píng)價(jià),其推廣性較差[81]。一個(gè)更直接更普通的評(píng)價(jià)方法是應(yīng)用認(rèn)知心理學(xué)理論,使用不同的可視化方法向被實(shí)驗(yàn)者展示問(wèn)題,通過(guò)評(píng)估回答的精確性、信心和花費(fèi)時(shí)間等項(xiàng)目比較不同可視化方法的績(jī)效[82]。
總體來(lái)說(shuō),信息可視化的評(píng)價(jià)方法和體系吸引了一些學(xué)者的注意,但是研究成果還是少量和滯后的,因此需要對(duì)信息可視化評(píng)價(jià)的理論基礎(chǔ)、方法和應(yīng)用做深入的研究[83]。
上述資料表明:信息可視化研究已經(jīng)取得了一些很有意義的成果,尤其是在可視化技術(shù)與可視化認(rèn)知方面積累了大量經(jīng)驗(yàn),也進(jìn)行了許多新的嘗試。根據(jù)信息可視化的十大原則[84],可視化技術(shù)與應(yīng)用還應(yīng)該繼續(xù)向以下4個(gè)方面努力:直觀化:直觀、形象地呈現(xiàn)數(shù)據(jù);關(guān)聯(lián)化:挖掘、突出呈現(xiàn)數(shù)據(jù)之間的關(guān)聯(lián);藝術(shù)化:增強(qiáng)數(shù)據(jù)呈現(xiàn)的藝術(shù)效果,符合審美規(guī)則;交互化:增強(qiáng)人機(jī)交互,實(shí)現(xiàn)即時(shí)數(shù)據(jù)操作。戴國(guó)忠等認(rèn)為數(shù)據(jù)可視化技術(shù)的發(fā)展方向是協(xié)同(collaboration)、分析過(guò)程(analytics)、計(jì)算(computational)和意會(huì)(sense-making)[85]。信息可視化未來(lái)研究方向可以包括以下幾個(gè)內(nèi)容。
1)信息可視化和數(shù)據(jù)挖掘的緊密結(jié)合 對(duì)于大數(shù)據(jù)的定義一直在不斷地增長(zhǎng)。現(xiàn)在網(wǎng)絡(luò)和很多企業(yè)的數(shù)據(jù)量都達(dá)到了TB級(jí)別。[86]可視化技術(shù)在處理海量數(shù)據(jù)時(shí)的速度和效率是值得關(guān)注的第一個(gè)問(wèn)題;第二問(wèn)題是傳統(tǒng)的可視化技術(shù)在面對(duì)多維度和大數(shù)據(jù)量時(shí)不可避免的造成視覺(jué)混亂(visual clutter)現(xiàn)象;非結(jié)構(gòu)化數(shù)據(jù),尤其是異構(gòu)的、非結(jié)構(gòu)化的混雜數(shù)據(jù)的出現(xiàn),對(duì)信息可視化提出了另一個(gè)大挑戰(zhàn),因此必須緊密結(jié)合信息可視化和數(shù)據(jù)挖掘技術(shù),運(yùn)用數(shù)據(jù)挖掘的公式和算法,對(duì)數(shù)據(jù)分析的過(guò)程及結(jié)果進(jìn)行可視化展現(xiàn),讓用戶可以方便高效的操作海量數(shù)據(jù),以發(fā)現(xiàn)隱含信息,從而引導(dǎo)出新的預(yù)見(jiàn)和更高效的決策。
2)協(xié)同可視化 由于海量數(shù)據(jù)的出現(xiàn)和技術(shù)方法的復(fù)雜性,信息可視化需要由多成員、多團(tuán)隊(duì)的開(kāi)發(fā)來(lái)完成,因此需要?jiǎng)?chuàng)造協(xié)同可視化環(huán)境,利用合理的數(shù)據(jù)分布使得多個(gè)工作站之間實(shí)現(xiàn)資源共享,對(duì)于可視化的過(guò)程進(jìn)行同步和異步控制,利用并行性來(lái)獲得極高的加速比,從而解決多個(gè)研究工作者間的協(xié)同性問(wèn)題。協(xié)同可視化領(lǐng)域的研究方向可以包括可視化接口設(shè)計(jì)、基于web的可視化協(xié)同平臺(tái)開(kāi)發(fā)、協(xié)同可視化工作的視圖設(shè)計(jì)、協(xié)同可視化中的工作流管理、及協(xié)同可視化技術(shù)的應(yīng)用等。
3)更多領(lǐng)域的應(yīng)用技術(shù) 信息可視化的方法和技術(shù)已經(jīng)應(yīng)用到了越來(lái)越多的領(lǐng)域,同時(shí)這些領(lǐng)域內(nèi)的應(yīng)用也為可視化技術(shù)的完善和發(fā)展提供良好的環(huán)境,當(dāng)前和未來(lái)信息可視化應(yīng)用的熱點(diǎn)領(lǐng)域包括以下4個(gè)方面。
統(tǒng)計(jì)可視化:需要研究使用幾何、動(dòng)畫(huà)、圖像等工具對(duì)數(shù)據(jù)統(tǒng)計(jì)的過(guò)程和結(jié)果進(jìn)行加工和處理的技術(shù)。
新聞可視化:通過(guò)對(duì)新聞內(nèi)容進(jìn)行抓取、清洗和提取和可視化展示,讓繁雜的信息有據(jù)可依、有線可尋、有圖可索,讓信息傳播更直觀、更生動(dòng)、更快捷。
社交網(wǎng)絡(luò)可視化:通過(guò)可視化方式顯示社交網(wǎng)絡(luò)的數(shù)據(jù),通過(guò)對(duì)社交網(wǎng)絡(luò)中節(jié)點(diǎn)、關(guān)系及時(shí)空數(shù)據(jù)的集成,有效揭示社交網(wǎng)絡(luò)的關(guān)聯(lián)、比較、走勢(shì)關(guān)系。
搜索日志可視化:針對(duì)在使用搜索引擎時(shí)產(chǎn)生的海量搜索日志,可視化的展現(xiàn)用戶的搜索行為、用戶與信息化環(huán)境的交互模式、用戶與用戶之間的交互關(guān)系,獲取用戶的需求和關(guān)系模式。
/
[1] DEFANTI T A, BROWN M D, MCCORMICK B H. Visualization:Expanding scientific and engineering research opportunities[J]. Computer, 1989,22(8) : 12-16.
[2] ROBERTSON G, CARD S K, MACKINLAY J D. The cognitive coprocessor architecture for interactive user interfaces[A]. Proceedings of the 2nd Annual ACM SIGGRAPH Symposium on User interface Software and Technology[C]. New York: ACM, 1989. 10-18.
[3] SINDIY O, LITOMISKY K, DAVIDOFF S, et al. Introduction to information visualization (InfoVis) techniques for Model-Based systems engineering[J]. Procedia Computer Science, 2013(16): 49-58.
[4] CARD S K, MACKINLAY J D, SHNEIDERMAN B. Readings in Information Visualization: Using Vision to Think[M]. San Francisco: Morgan Kaufmann, 1999.
[5] 陳建軍,于志強(qiáng),朱 昀. 數(shù)據(jù)可視化技術(shù)及其應(yīng)用[J]. 紅外與激光工程, 2001,30(5): 339-342. CHEN Jianjun, YU Zhijiang, ZHU Yun. Data visualization and its applications[J]. Infrared and Laser Engineering, 2001,30(5):339-342.
[6] 任 磊,王威信,周明駿,等. 一種模型驅(qū)動(dòng)的交互式信息可視化開(kāi)發(fā)方法[J]. 軟件學(xué)報(bào),2008,19(8): 1947-1967. REN Lei, WANG Weixin, ZHOU Mingjun, et al. A model driven development method for interactive information visualization[J]. Journal of Software, 2008, 19(8):1947-1967.
[7] 劉大海. 海量數(shù)據(jù)可視化方法的研究[D]. 天津:天津大學(xué),2009. LIU Dahai. The Research of Large-scale Data Visualization[D]. Tianjin: Tianjin University, 2009.
[8] SHNEIDERMAN B. The eyes have it: A task by data type taxonomy for information visualizations[A]. Proceedings of the 1996 IEEE Symposium on Visual Languages[C]. Washington DC: IEEE Computer Society, 1996. 336-343.
[9] INSELBERG A. The plane with parallel coordinates[J]. The Visual Computer, 1985, 1(2):69-91.
[10] 孫 揚(yáng),封孝生,唐九陽(yáng),等.多維可視化技術(shù)綜述[J].計(jì)算機(jī)科學(xué),2008, 35(11):1-7. SUN Yang, FENG Xiaosheng, TANG Jiuyang, et al. Survey on the research of multidimensional and multivariate data visualization[J]. Computer Science, 2008, 35(11): 1-7.
[11] INSELBERG A, DIMSDALE B. Parallel coordinates[A]. Human-Machine Interactive Systems[C]. Berlin: Springer, 1991.199-233.
[12] 楊 峰,李月華.高維信息可視化方法研究綜述[J].情報(bào)理論與實(shí)踐,2012, 35(9):125-128. YANG Feng, LI Yuehua. Review of multidimensional information visualization[J]. Information Studies: Theory & Application, 2012, 35(9): 125-128.
[13] FUA Y H, WARD M O, RUNDENSTEINER E A. Hierarchical parallel coordinates for exploration of large datasets[A]. Proceedings of the Conference on Visualization’99: Celebrating Ten Years[C]. Los Alamitos: IEEE Computer Society Press, 1999.443-508.
[14] GRAHAM M, KENNEDY J. Using curves to enhance parallel coordinate visualizations[A]. Proceedings of the Seventh International Conference on Information Visualization[C]. Washington DC: IEEE Computer Society, 2003. 10-16.
[15] 盛秀杰,金之鈞,王義剛.一種新的面向多元統(tǒng)計(jì)分析的信息可視化技術(shù)[J].石油地球物理勘探,2013, 48(3):488-496. SHENG Xiujie, JIN Zhijun, WANG Yigang. A novel information visualization approach for multiple statistical analysis[J]. Oil Geophysical Prospecting, 2013, 48(3): 488-496.
[16] SIIRTOLA H. Direct manipulation of parallel coordinates[A]. CHI’00 Extended Abstracts on Human Factors in Computing Systems[C]. New York: ACM, 2000. 373-378.
[17] WONG P C, BERGERON R D. Multiresolution multidimensional wavelet brushing[A]. Proceedings of the 7th Conference on Visualization’96[C]. Los Alamitos: IEEE Computer Society Press, 1996. 141-148.
[18] HAUSER H, LEDERMANN F, DOLEISCH H. Angular brushing of extended parallel coordinates[A]. IEEE Symposium on Information Visualization 2002[C].Washington DC: IEEE Computer Society, 2002. 127-130.
[19] WONG N, CARPENDALE S, GREENBERG S. Edge lens: An interactive method for managing edge congestion in graphs[A]. Proceedings of the IEEE Symposium on Information Visualization (InfoVis’02)[C]. Washington DC: IEEE Computer Society, 2002. 51-58.
[20] ZHOU H, YUAN X R, QU H M, et al. Visual clustering in parallel coordinates[J]. Computer Graphics Forum, 2008, 27(3): 1047-1054.
[21] BERTINI E, DELL L, SANTUCCI G. SpringView: Cooperation of radviz and parallel coordinates for view optimization and clutter reduction[A]. Proceedings of the Coordinated and Multiple Views in Exploratory Visualization[C].Washington DC: IEEE Computer Society, 2005.22-29.
[22] YANG J, PENG W, WARD M O, et al. Interactive hierarchical dimension ordering, spacing and filtering for exploration of high dimensional datasets[A]. Proceedings of the Ninth annual IEEE Conference on Information visualization[C].Washington DC: IEEE Computer Society, 2003. 105-112.
[23] JOHANSSON J, LJUNG P, JERN M, et al. Revealing structure within clustered parallel coordinates displays[A]. Proceedings of the 2005 IEEE Symposium on Information Visualization[C]. Washington DC: IEEE Computer Society, 2005. 125-132.
[24] JOHANSSON J, COOPER M. A screen space quality method for data abstraction[J]. Computer Graphics Forum, 2008, 27(3):1039-1046.
[25] SHEARER J, OGAWA M, KWAN-LIU M, et al. Pixelplexing: Gaining display resolution through time[A]. IEEE Pacific Visualization Symposium[C]. Washington DC: IEEE Computer Society, 2008. 159-166.
[26] KANDOGAN E. Visualizing multi-dimensional clusters, trends, and outliers using star coordinates[A]. Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining[C]. New York: ACM, 2001. 107-116.
[27] CRAIG P, KENNEDY J. Coordinated graph and scatter-plot views for the visual exploration of microarray time-series data[A]. Proceedings of the Ninth Annual IEEE Conference on Information Visualization[C].Washington DC: IEEE Computer Society, 2003. 173-180.
[28] SCHMID C, HINTERBERGER H. Comparative multivariate visualization across conceptually different graphic displays[A]. Proceedings of the Seventh International Working Conference on Scientific and Statistical Database Management[C]. Washington DC: IEEE Computer Society, 1994. 42-51.
[29] ANDREWS D F. Plots of high dimensional data[J]. International Journal of Biometrics, 1972, 28(1): 125-136.
[30] CHERNOFF H. The use of faces to represent points in k-dimensional space graphically[J]. Journal of the American Statistical Association, 1973(68): 361-368.
[31] LEE M D, BUTAVICIUS M A, REILLY R E. Visualizations of binary data: A comparative evaluation[J]. International Journal of Human-computer Studies, 2003, 59(5): 569-602.
[32] ROBERTSON G G, MACKINLAY J D, CARD S K. Cone trees: animated 3D visualizations of hierarchical information[A]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems[C]. New York: ACM, 1991.189-194.
[33] HEER J, ROBERTSON G G. Animated transitions in statistical data graphics[J]. IEEE transaction on Visualization and Computer Graphics, 2007, 13(6): 1240-1247.
[34] ROBERTSON G, CAMERON K, CZERWINSKI M, et al. Animated visualization of multiple intersecting hierarchies[J]. Journal of Information Visualization, 2002, 1(1): 50-65.
[35] PLAISANT C, GROSJEAN J, BEDERSON B B. Space tree: Supporting exploration in a large node-link tree, design evolution and empirical evaluation[A]. Proceedings of the IEEE Symposium on Information Visualization (InfoVis’02)[C].Washington DC: IEEE Computer Society, 2002.57-64.
[36] HEER J, CARD S K. DOITrees revisited: Scalable, space-constrained visualization of hierarchical data[A]. Proceedings of the Working conference on Advanced Visual Interfaces[C]. New York: ACM, 2004. 421-424.
[37] WATTENBERG M, KRISS J. Designing for social data analysis[J]. IEEE Trans on Visualization and Computer Graphics, 2006,12(4):549-557.
[38] BUSH S,MENZIES G,THORP S.An array of online teaching tools[J].Teaching Statistics,2009,31(1):17-20.
[39] TUFTE E. The Visual Display of Quantitative Information[M]. Cheshire: Graphics Press, 2001.
[40] SAITO T, MIYAMURA H N, YAMAMOTO M, et al. Two-tone pseudo-coloring: Compact visualization for one-dimensional data[A]. Proceedings of the 2005 IEEE Symposium on Information Visualization[C]. Washington DC: IEEE Computer Society, 2005. 173-180.
[41] HEER J, KONG N, AGRAWALA M. Sizing the horizon: The effects of chart size and layering on the graphical perception of time series visualizations[A]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems 2009[C]. New York: ACM, 2009. 1303-1312.
[42] KIM N W, CARD S K, HEER J. Tracing genealogical data with TimeNets[A]. Proceedings of the International Conference on Advanced Visual Interfaces[C]. New York: ACM, 2010.26-28.
[43] SPENCE R, APPERLEY M D. Database navigation: An office environment for the professional[J]. Behavior and Information Technology, 1982(1): 43-54.
[44] MACKINLAY J D, ROBERTSON G G, AND CARD S K. The perspective wall: Detail and context smoothly integrated[A]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems[C]. New York: ACM, 1991.173-176.
[45] JENSEN M. Visualizing Complex Semantic Timelines[R]. [S.l.]:[s.n.],2003.
[46] ROBERTSON G G, MACKINLAY J D, CARD S K. Cone trees:Animated 3D visualizations of hierarchical information[A]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems[C]. New York: ACM, 1991.189-194.
[47] 張 昕,袁曉如.樹(shù)圖可視化[J].計(jì)算機(jī)輔助設(shè)計(jì)與圖形學(xué)學(xué)報(bào),2012, 24(9):1113-1124. ZHANG Xin, YUAN Xiaoru. Treemap visualization[J]. Journal of Computer-aided Design & Computer Graphics, 2012, 24(9): 1113-1124.
[48] CARD S K, SUH B, PENDLETON B A, et al. Time tree: Exploring time changing hierarchies[A]. 2006 IEEE Symposium on Visual Analytics Science and Technology[C].Washington DC: IEEE Computer Society, 2006. 3-10.
[49] SELASSIE D, HELLER B, HEER J. Divided edge bundling for directional network data[J]. IEEE Trans Visualization and Computer Graphics,2011,17(12):2354-2363.
[50] BERTIN J. Sémiologie Graphique[M]. Paris:University of Wisconsin Press, 1983.
[51] CLEVELAND W S, MCGILL R. Graphical perception: Theory, experimentation, and application to the development of graphical methods[J]. Journal of the American Statistical Association, 1984, 387 (79):531-554.
[52] SIMKIN D, HASTIE R. An information-processing analysis of graph perception[J]. Journal of the American Statistical Association, 1987, 398 (82):454-465.
[53] TALBOT J, LIN S, HANRAHAN P. An extension of Wilkinson's algorithm for positioning tick labels on axes[J]. IEEE Transactions on Visualization and Computer Graphics, 2010, 16(6): 1036-1043.
[54] HEER J, STONE M. Color naming models for color selection, image editing and palette design[A]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems[C]. New York: ACM, 2012. 1007-1016.
[55] BERGMAN L D, ROGOWITZ B E, TREINISH L A. A rule-based tool for assisting colormap selection[A]. Proceedings of the 6th conference on Visualization’95[C]. New York: ACM, 1995.118-125.
[56] HARROWER M, BREWER C.Colorbrewer org:An online tool for selecting color schemes for maps[J]. The Cartographic Journal, 2003, 40(1): 27-37.
[57] MEIER B J, SPALTER A M, KARELITZ D B. Interactive color palette tools[J]. IEEE Computer Graphics and Applications, 2004, 24(3): 64-72.
[58] COHEN-OR D, SORKINE O, GAL R. Color harmonization[J]. ACM Transactions on Graphics, 2006, 25(3):624-630.
[59] WANG L J, GIESEN J, MCDONNELL K T, et al. Color design for illustrative visualization[J]. Visualization and Computer Graphics, 2008, 14(6): 1739-1746.
[60] CHUANG J, WEISKOPF D, MOELLER T, et al. Energy aware color sets[J]. Computer Graphics Forum, 2009, 28(2): 203-211.
[61] LIN S, FORTUNA J, KULKARNI C, et al. Selecting semantically-resonant colors for data visualization[J]. Computer Graphics Forum, 2013,32(3).
[62] ROUSSINOV D, RAMSEY M. Information forage through adaptive visualization[A]. Proceedings of the Third ACM conference on Digital Libraries[C]. New York: ACM, 1998. 303-304.
[63] DOMIK G O, GUTKAUF B. User modeling for adaptive visualization systems[A]. Proceedings of the conference on Visualization’94[C]. Los Alamitos: IEEE Computer Society Press, 1994. 217-223.
[64] GRAWEMEYER B, COX R. A bayesian approach to modelling users' information display preferences[A]. Proceedings of the 10th International conference on User Modeling[C]. Berlin: Springer, 2005. 225-230.
[65] TERAOKA T, MARUYAMA M. Adaptive information visualization based on the user's multiple viewpoints-interactive 3D visualization of the WWW[A]. Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis ’97)[C]. Washington DC: IEEE Computer Society,1997.25-28.
[66] BITTON E. A spatial model for collaborative filtering of comments in an online discussion forum[A]. 3rd ACM Conference on Recommender Systems[C]. New York: ACM, 2009. 393-396.
[67] GANSNER E, HU Y, KOBOUROV S, et al. Putting recommendations on the map: Visualizing clusters and relations[A]. Proceedings of the Third ACM Conference on Recommender Systems[C]. New York: ACM, 2009.345-348.
[68] BRUSILOVSKY P, RIZZO R. Using maps and landmarks for navigation between closed and open corpus hyperspace in web-based education[J]. New Review of Hypermedia and Multimedia, 2003, 8(1): 59-82.
[69] MEYER J, GELDER S, KRETSCHMER K, et al. Interactive visualization of hybrid medical data sets[A]. Proceeding of WSCG[C]. Plzen: University of West Bohemia Press, 1997. 371-380.
[70] GALLOP J R, BLAIR G S, COOPER C S, et al. Component-oriented approach to adaptive collaborative visualization[A]. Visualization and Data Analysis[C]. Bellingham: SPIE, 2002. 99-104.
[71] LIU H H, JI Q G. Research on adaptive rendering based on visual acuity equations[A]. Fourth International Conference on vVirtual Reality and Its Applications in Industry[C]. Bellingham: SPIE, 2004. 96-99.
[72] BRUSILOVSKY P. Methods and techniques of adaptive hypermedia[J]. User Modeling and User-adapted Interaction, 1996, 6(2/3): 87-129.
[73] WIZA W, WALCZAK K, CELLARY W. Periscope: A system for adaptive 3D visualization of search results[A]. Proceedings of the Ninth International Conference on 3D Web Technology[C]. New York: ACM, 2004. 29-40.
[74] BAKALOV F, K?NIG-RIES B, NAUERZ A, et al. IntrospectiveViews: An interface for scrutinizing semantic user models[A]. Proceedings of the 18th International Conference on User Modeling, Adaptation, and Personalization[C].Berlin: Springer, 2010.219-230.
[75] KUMP B, SEIFERT C, BEHAM G, et al. MyExperiences:Visualizing evidence in an open learner model[A]. Adjunct Proceedings of the 18th International Conference on User Modeling, Adaptation, and Personalization[C].Big Island HI:[s.n.], 2010.16-18.
[76] AHN J W, BRUSILOVSKY P, GRADY J, et al. Open user profiles for adaptive news systems: Help or harm[A]. Proceedings of the 16th International Conference on World Wide Web[C]. New York:ACM,2007.11-20.
[77] MORSE E, LEWIS M, OLSEN K A. Evaluating visualizations: Using a taxonomic guide[J]. International Journal of Human-computer Studies, 2000, 53(5): 637-662.
[78] MEYER J. Performance with tables and graphs: Effects of training and a visual search model[J]. Ergonomics, 2000, 43(11): 1840-1865.
[79] AHN J W, BRUSILOVSKY P. Adaptive visualization for exploratory information retrieval[J]. Information Processing & Management, 2013, 49(5): 1139-1164.
[80] FREITAS C M D S, LUZZARDI P R G, CAVA R A, et al. On evaluating information visualization techniques[A]. Proceedings of the Working Conference on Advanced Visual Interface[C]. New York: ACM, 2002.373-374.
[81] DOWELL J, LONG J. Conception of the cognitive engineering design problem[J]. Ergonomics, 1998, 41(2): 126-139.
[82] PURCHASE H C. The effects of graph layout[A]. Proceedings of the Australasian Conference on Computer Human Interaction[C]. Washington DC: IEEE Computer Society, 1998. 80-86.
[83] 楊 峰,李 蔚.評(píng)價(jià)信息可視化技術(shù)的指標(biāo)研究[J].圖書(shū)情報(bào)知識(shí),2007(4):80-84. YANG Feng, LI Wei. The research of indexes for evaluating information visualization techniques[J]. Document, Information & Knowledge, 2007(4): 80-84.
[84] KELLEHER C, WAGENER T. Ten guidelines for effective data visualization in scientific publications[J]. Environmental Modelling & Software , 2011,26:822-827.
[85] 戴國(guó)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[J].中國(guó)科學(xué),2013, 43(1):178-184. DAI Guozhong, CHEN Wei, HONG Wenxue, et al. Information visualization and visual analytics: Challenges and opportunities[J]. China Science, 2013, 43(1): 178-184
[86] 任永功,于 戈.?dāng)?shù)據(jù)可視化技術(shù)的研究與進(jìn)展[J].計(jì)算機(jī)科學(xué),2004, 31(12):92-96. REN Yonggong, YU Ge. Research and development of the data visualization techniques[J]. Computer Science, 2004, 31(12): 92-96.
Review of information visualization
YANG Yanbo1, LIU Bin1, QI Mingyue2
(1.School of Economics and Management, Hebei University of Science and Technology, Shijiazhuang Hebei 050018, China;2.Communication Station of Hebei Provincial Military Command, Shijiazhuang Hebei 050011, China)
Information visualization is the application of visualization technology in non-spatial data area, enhancing data presentation effect.Users can observe the data intuitively and interactively so as to find implicit features, relations and patterns in data. The application of information visualization is very abroad which includes data mining visualization, network data visualization, social data visualization, traffic visualization, text visualization, and medicine visualization, etc. According to Card model on information visualization, process of information visualization includes three stages:data pretreating, data plotting, data displaying and interacting.Ben Shneiderman notes that visualization data includes one-dimensional data, two-dimensional data, three-dimensional data, multi-dimensional data, temporal data, hierarchical data, and network data, of which are given much attention to research.
Visualization methods of multi-dimensional data include geometry methods, icon methods, and animation methods, etc. Among the geometry-based visualization methods, the most classic one is the parallel coordinates approach. It uses parallel vertical axis to represent the dimension values. By the multidimensional data portrayed on the shaft, and by the coordinate point connected with a line to a data entry on the axes, the multidimensional data was presented. Multi-dimensional data was displayed concisely and quickly in Parallel Coordinates, and improved many techniques. When scale of data set was very large, the dense lines could cause visual clutter. The methods of clutter reduction include dimension reordering, interacting, clustering and filtering, and visual enhancement, etc. Other methods based on geometry, including Radviz (Radial Coordinate visualization), display multi-dimensional data by circular coordinate. Scatter plot matrix arranges every demensions of multidimensional data to be combined into pairwise mode, drawing a series of regular scatters. Icon was used to describe the multi-dimensional data by its geometrical features including size, length, form and color, etc. Icon methods include star graph and Chernoff face method. Animation used for visualization can improve the degree of interacting and understanding. , but with shortcomings such as:distraction, misunderstanding and visual clutter.
Time serial data refers to data sets with time property. The visualization methods include line chart, stock chart, animation, horizon graph and Timeline.
Hierarchical data can be used to describe object whose attributes are rank and level. Its visualization methods include linking point graph and tree map. Tree map displays hierarchical data by nesting hoop and lump. For displaying more content, based on "Focus+Content" technology, some methods were put forward including "fish eyes" technology, geometry deformation, Semantic zooming and clustering.Network data has network structure. Layout algorithm is the core of visualization of network data, which includes three classes:Force-Directed Layout, Hierarchical Layout and Grid Layout. When there're many data connection nodes, edge corssover phenomenon happens, causing visual confusion. There were a variety of techniques for resolving the edge bundling, including hierarchical edge bundling, force-directed edge bundling, geometry-based edge clustering, multi-level agglomerative edge bundling, and grid-based methods.
Other research hotspots include research on visual feature, adaptive visualization and evaluation of information visualization.
Effect of visual feature such as position, length, area, shape and color, etc. on visual result has received considerable attention. Color is one of most important visual factor, so research focuses on the color selection principle and interaction system, which are based on data type, quantity, and cognitive constraints.
Adaptive visualization can enhance adaptability of information visualization, which includes adaptive display, adaptive resource model, and adaptive user model according to research of Domik & Gutkauf and Grawemeyer & Cox. Adaptive display provides automatic and suitable display for different users, including selecting content and layout, adjusting visual features automatically. Adaptive resource model means utilizing hardware and software to enhance visual performance. Adaptive user model means displaying user model in order to edit and control content.
Morse et al. notes that the research on evaluation of information visualizations is rare. Evaluation on direct and general information visualization was not involved in some research. So, it is needed to do deep research on the theoretical basis, method and application of information visualization evaluation.
Technology and application of information visualization should be developed in four aspects, displaying data directly perceived through the senses, mining and showing relation between data, strengthening demonstration of aesthetics and artistry, enhancing performance of interaction and operation on real-time data. Dai et al. noted that research direction of information visualization was Collaboration, Analytics, Computational and Sense-making. Research directions in future is as following.
Visualization and data mining:to promote efficiency and avoid visual clutter in processing huge data, information visualization should be combined with data mining so that user can operate huge data and discover implicit information.
Collaborative visualization:Collaborative visualization includes interface design, collaborative platform based on web, view design, workflow design, and application of technology.
Application in more fields:statistics visualization refers to processing and handling the statistical process data and results by method of geometry, animation, and graph ett. News visualization refers to presenting diversely analysis results after grasping, cleaning, and drawing news corpus. Social network visualization refers to displaying and revealing relation, comparison, and trend of social network through integration of dimensions of time and space. Search log visualization refers to displaying huge searching behavior when using a search engine. Users' search behavior, relationships and patterns are presented visually.
information visualization; visualization technology; human-machine interaction; data mining
1008-1542(2014)01-0091-12
10.7535/hbkd.2014yx01016
2013-10-16;
2013-11-20;責(zé)任編輯:李 穆
國(guó)家自然科學(xué)基金(71271076)
楊彥波(1981-),男,河北任縣人,講師,博士,主要從事數(shù)據(jù)挖掘、信息可視化、商務(wù)智能方面的研究。
E-mail:yangyb7@163.com
TP391
A
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.