搜索
BDTC-Beijing,2013-12-6BDTC-Beijing,2013-12-6BigDataVisualizationandVisualAnalysis-theChallengesandOpportunities ⏠⏠BDTC-Beijing,2013-12-6Visualization3BDTC-Beijing,2013-12-6! :…! ! ·! ·4BDTC-Beijing,2013-12-6! (Visualization)(mentalmodel)(mentalimage)5VisualizationDataMentalModelImageInsightsBDTC-Beijing,2013-12-6FromDatatoVisualization6BDTC-Beijing,2013-12-6Visualization!=Infographics! 7BDTC-Beijing,2013-12-68BDTC-Beijing,2013-12-6LargeVolumeVisualization! LevelofDetails! OutofCore! ParallelVisualization9BDTC-Beijing,2013-12-6Top10ChallengesinExtreme-ScaleDataVisualAnalytics10PakChungWong(PNNL)Han-WeiShen(OSU)ChrisJohnson(Utah)ChaomeiChen(Drexel)RobertRoss(Argonne)BDTC-Beijing,2013-12-6Top10ChallengesinExtreme-ScaleDataVisualAnalytics! InSituAnalysis! Performasmuchanalysisaspossiblewhilethedataarestillinmemory! InteractionandUserInterfaces! Machine-basedautomatedsystemsvs.HumanCognition! Large-DataVisualization! DataprojectionanddimensionReduction,displaytechnology! DatabasesandStorage! Acloud-basedsolutionmightnotmeettheneeds! Algorithms! Addressbothdata-sizeandvisual-efficiencyissues11BDTC-Beijing,2013-12-6Top10ChallengesinExtreme-ScaleDataVisualAnalytics! DataMovement/Transport,&NetworkInfrastructure! Efficientlyusenetworkingresourcesandprovideconvenientabstractions! UncertaintyQuantification! Copewithincompletedata! Parallelism! DomainandDevelopmentLibraries,Frameworks,andTools! Affordableresourcelibraries,frameworks,andtools! Social,Community,andGovernmentEngagements12BDTC-Beijing,2013-12-6ChallengesinBigDataVisualization/VisualAnalytics-1! IntegratingheterogeneousDatafromdifferentresourcesandscales13BDTC-Beijing,2013-12-6BeijingTaxiGPSdata14BDTC-Beijing,2013-12-6Data! BeijingtaxiGPSdata! Size:34.5GB! Taxinumber:28,519! Samplingpointnumber:379,107,927! Timerange:2009/03/02~25(24days,but03/18dataismissing)! Samplingrate:30secondsperpoint(but60%datamissing)! Beijingroadnetwork(fromOpenStreetMap)! Size:40.9MB! 169,171nodesand35,422ways15BDTC-Beijing,2013-12-616BDTC-Beijing,2013-12-617BDTC-Beijing,2013-12-6TrafficJamDetection18Traffic&Jam&Detec,onRaw&taxi&GPS&DataRaw&Road&NetworkCleaned&GPS&DataProcessed&&Road&NetworkGPS&Trajectories&Matched&to&the&Road&NetworkRoad&Speed&DataTraffic&Jam&Event&Data………9:10&am55&km/h9:20&am10&km/h9:30&am12&km/h9:40&am45&km/h………aDefining&propaga,on&based&on&spa,al/temporal&rela,onship:&be0e1………9:10&am50&km/h9:20&am45&km/h9:30&am12&km/h9:40&am15&km/h………e0&happens&before&e1,&and&on&a&dWay&following&e1e1e0BDTC-Beijing,2013-12-6VisualInterface19Road&Speed&DataTraffic&Jam&Event&DataTraffic&Jam&Propaga,on&GraphsPropaga,on&Graphs&of&InterestOne&Propaga,on&GraphRoad&Segment&Level&Explora,on&and&AnalysisPropaga,on&Graph&Level&Explora,onTime&and&Size&Distribu,onSpa,al&DensityPropaga,on&Graph&ListSpa,al&FilterTemporal&&&Size&FilterTopological&FilterTopological&Clustering&&Road&of&InterestDynamic&QueryBDTC-Beijing,2013-12-6Preprocessing:MapMatching20MapMatchingRawtaxiGPSDataRawRoadNetworkCleanedGPSDataProcessedRoadNetworkGPSTrajectoriesMatchedtotheRoadNetworkBDTC-Beijing,2013-12-621BDTC-Beijing,2013-12-6VisualInterface:SingleRoadLevel! Pixelbasedvisualization22Timeofaday:144columns(eachfora10min)Days:24rows(eachforoneday)EachcellrepresentsonetimebinColorencodespeedBDTC-Beijing,2013-12-6CaseStudy:RoadLevelExplorationandAnalysis! Differentroadcongestionpatterns23BDTC-Beijing,2013-12-6CaseStudy:RoadLevelExplorationandAnalysis24BDTC-Beijing,2013-12-6PropagationGraphAnalysis! SpatialTemporalinformationofonepropagation25SpatialpathTemporaldelayLargedelayBDTC-Beijing,2013-12-6PropagationPatternExploration! Propagationgraphsforoneregioninthemorningofdifferentdays26BDTC-Beijing,2013-12-6! 27BDTC-Beijing,2013-12-628BDTC-Beijing,2013-12-629BDTC-Beijing,2013-12-630BDTC-Beijing,2013-12-631BDTC-Beijing,2013-12-632BDTC-Beijing,2013-12-633BDTC-Beijing,2013-12-634BDTC-Beijing,2013-12-635BDTC-Beijing,2013-12-6WeiboThemeMap36BDTC-Beijing,2013-12-6WeiboThemeMap! 37BDTC-Beijing,2013-12-6XiamenTraffic38BDTC-Beijing,2013-12-6! 39BDTC-Beijing,2013-12-6ChallengesinBigDataVisualization/VisualAnalytics-2! IntegratingheterogeneousDatafromdifferentresourcesandscales! ScalabilityinData/Taskcomplexity! Datainherentpropertiesimposemorecomputationalchallengesmethodsforvisualizationandvisualanalysisonbigdata40BDTC-Beijing,2013-12-6PollutionFromChina41BDTC-Beijing,2013-12-6PollutionfromUSA42BDTC-Beijing,2013-12-6MultivariatetoMulti-RunVisualAnalysis43QVAPORQCLOUDSpeedPressureQVAPORQVAPORQCLOUDSpeedPressureQVAPORQCLOUDSpeedPressureQVAPORQCLOUDSpeedPressure(Multivariate)(EnsembleRuns)Run1Run2Run3BDTC-Beijing,2013-12-6EulerianandLagriangianSpecifications! Eulerian:! Lagriangian:! Relationshipsbetweentwospecifications(flowmap):44BDTC-Beijing,2013-12-6Eulerian-basedAttributeSpaceProjection! Samplesondatagrid!Samplesinattributespace!Eulerian-basedAttributeSpaceProjection(EASP)45BDTC-Beijing,2013-12-6Lagrangian-basedAttributeSpaceProjection! Pathlinesondatagrid!Pathlinesinattributespace!Lagrangian-basedAttributeSpaceProjection(LASP)! Both