Interactive Exploration of Multidimensional DataIs It Really That Common ?I can’t see it, I tell ya !Standard SolutionBut there are so many !Introducing Rank-by-featureHow does it work ?Ranking Criteria - 1DRanking Criteria - 2DPut A Demo Where Your Mouth Is !PowerPoint PresentationSlide 12Slide 13Slide 14Slide 15Questions/Critiques1Interactive Exploration of Multidimensional DataInteractive Exploration of Multidimensional DataBy:Sanket SinhaNitin MadnaniBy:Sanket SinhaNitin Madnani2Is It Really That Common ?Is It Really That Common ?You Bet:DemographicsEconomicsCensusMicroarray Gene ExpressionEngineeringPsychologyHealthYou Bet:DemographicsEconomicsCensusMicroarray Gene ExpressionEngineeringPsychologyHealth3I can’t see it, I tell ya !I can’t see it, I tell ya !Visualization challenges for >= 3D:Relationship comprehension is difficultDiscovering outliers, clusters and gaps is almost impossibleOrderly exploration is not possible with standard visualization systemsNavigation is cognitively onerous and disorienting (3D)Occlusion (3D)Visualization challenges for >= 3D:Relationship comprehension is difficultDiscovering outliers, clusters and gaps is almost impossibleOrderly exploration is not possible with standard visualization systemsNavigation is cognitively onerous and disorienting (3D)Occlusion (3D)4Standard SolutionStandard SolutionCan you say “Pro-jek-shun” ?Use lower dimensional projections of data:Can you say “Pro-jek-shun” ?Use lower dimensional projections of data:1D : Histograms2D : Scatterplots5But there are so many !But there are so many !For 13 dimensions (columns) : Number of histograms = 13Number of scatterplots = C(13,2) = 78Must examine a series of these to gain insightsUnsystematic == InefficientMust have order !For 13 dimensions (columns) : Number of histograms = 13Number of scatterplots = C(13,2) = 78Must examine a series of these to gain insightsUnsystematic == InefficientMust have order !6Introducing Rank-by-featureIntroducing Rank-by-featureAllows projections to be examined in an orderly fashionA powerful framework for interactive detection of:Inter-dimension relationshipsGapsOutliersPatternsAllows projections to be examined in an orderly fashionA powerful framework for interactive detection of:Inter-dimension relationshipsGapsOutliersPatterns7How does it work ?How does it work ?Framework defines ranking criteria for 1D & 2D projectionsUser selects criterion of interestAll projections are scored on the criterion and rankedUser examines projections in the order recommendedEureka* !!Framework defines ranking criteria for 1D & 2D projectionsUser selects criterion of interestAll projections are scored on the criterion and rankedUser examines projections in the order recommendedEureka* !!*Disclaimer: All users may not be able to make life-altering discoveries8Ranking Criteria - 1DRanking Criteria - 1DNormality: Indicative of how “Gaussian” the dataset isUniformity: How “uniform” is the dataset ?(How high is the entropy ?)Outliers: The number of potential outliers in the datasetGap: The size of the biggest gapUniqueness: Number of unique data pointsNormality: Indicative of how “Gaussian” the dataset isUniformity: How “uniform” is the dataset ?(How high is the entropy ?)Outliers: The number of potential outliers in the datasetGap: The size of the biggest gapUniqueness: Number of unique data points9Ranking Criteria - 2DRanking Criteria - 2DLinear Correlation: Pearson’s correlation coefficientLSE: Least Square Error from the optimal quadratic curve fitQuadracity: Quadratic coefficient from fitting curve equationUniformity: Joint entropyROI: Number of items in a Region Of InterestOutliers: Number of potential outliersLinear Correlation: Pearson’s correlation coefficientLSE: Least Square Error from the optimal quadratic curve fitQuadracity: Quadratic coefficient from fitting curve equationUniformity: Joint entropyROI: Number of items in a Region Of InterestOutliers: Number of potential outliers10Put A Demo Where Your Mouth Is !Put A Demo Where Your Mouth Is !11HCE OverviewHCE Overview12The Input Dialog BoxThe Input Dialog BoxPerform Filtering & NormalizationPerform Filtering & Normalization13Histogram OrderingHistogram Ordering14Scatterplot OrderingScatterplot Ordering15Tabular View of DataTabular View of DataSelect specific data records and annotate if neededSelect specific data records and annotate if needed16Questions/CritiquesQuestions/CritiquesWhat does “outlierness” mean?Cannot identify datapoints in histogram or scatterplot browser without switching to table viewEspecially in ROIHow to intuitively interpret:Outliers in 2DLSEQuadracityWhat does “outlierness” mean?Cannot identify datapoints in histogram or scatterplot browser without switching to table viewEspecially in ROIHow to intuitively interpret:Outliers in