CS 796/896 Visual Analytics Seminar Spring 2011 What is Visual Analytics? Dr. Michele C. Weigle http://www.cs.odu.edu/~mweigle/cs796-s11/ What is Visual Analytics?! New multidisciplinary field! Combines various research areas! visualization! human-computer interaction! data analysis! data management! geo-spatial and temporal data processing! spatial decision support! statistics http://www.youtube.com/watch?v=k9pvskathgi 2
Information Overload Problem! The danger of getting lost in data, which may be! irrelevant to the current task at hand! processed in an inappropriate way, or! presented in an inappropriate way! There is a need for methods and models to turn the data into reliable and comprehensible knowledge 3 Goal of Visual Analytics! Who or what defines the "relevance of information" for a given task?! How can inappropriate procedures in a complex decision making process be identified?! How can the resulting information be presented in a decision-oriented or task-oriented way? 4
The Need for Visual Analytics! Essential in application areas where large information spaces have to be processed and analyzed! astrophysics! monitoring climate and weather! emergency management! network security! terrorism informatics! biology and medicine! business intelligence! web science digital libraries (WS-DL) 5 Visualization in WS-DL Papers (Jan 25-Feb 1)! Zoetrope: Interacting with the Ephemeral Web! "Catch Me If You Can": Visual Analysis of Coherence Defects in Web Archiving! VisGets: Coordinated Visualizations for Webbased Information Exploration and Discovery! Describing Story Evolution from Dynamic Information Streams 6
Visual Analytics in Action Simulation of Climate Models 7 Visual Analytics in Action Distributed Network Attack on SSH 8 Tool: CGV [113] Tool: NFlowVis [76]
Visual Analytics in Action Cooling Jacket Simulation Tool: SimVis 9 Visual Analytics in Action Multivariate Datasets in Demography 10
But, It's Not Just For Massive Data Multi-dimensional analysis of 738- row spreadsheet Problem is not number of observations, but number of dimensions Note: next few slides present viewpoint of the CEO of Tableau Software 11 From Problems Chabot, with "Demystifying Visual Analytics, Keim Visual et al. Analytics", 2009 And, It Doesn't Have To Be Complex! Many people adopt visual analytics to help them see and understand relatively simple problems.! Standard visual paradigms (lines, bars, maps) are often better at representing data than fancy new visualizations. 12 From Problems Chabot, with "Demystifying Visual Analytics, Keim Visual et al. Analytics", 2009
13 From Problems Chabot, with "Demystifying Visual Analytics, Keim Visual et al. Analytics", 2009 What's The Goal?! People spend lots of time with data! exploring it! cleaning it! gaining confidence in it! summarizing it! confirming facts! presenting findings! Make this fast!! let people apply computing operations on data by interacting directly with visual representations 14 From Problems Chabot, with "Demystifying Visual Analytics, Keim Visual et al. Analytics", 2009
The Visual Analytics Process! Combines automatic and visual analysis methods! Tight coupling through human interaction! Goal: gain knowledge from data 15 The Visual Analytics Process 16
Building Blocks of Visual Analytics Research 17 Visualization! Scientific visualization! 3D data! real-world, physical models! Information visualization! abstract data! no explicit spatial references! data values cannot be naturally mapped to 2D or 3D display space! interaction is important 18
Data Management! Integration of heterogeneous data! Data cleansing! dealing with missing, inaccurate values, inconsistent labels! Google Refine - http://code.google.com/p/google-refine/! Makes use of intelligent data analysis techniques and visualization! Challenges:! streaming data sources! automatic extraction of information from large document collections (i.e., the web) 19 Data Management Papers (Feb 8-15)! Duplicate Record Detection: A Survey! A Taxonomy of Clutter Reduction for Information Visualisation! Clutter Reduction in Multi-Dimensional Data Visualization Using Dimension Reordering! Give Chance a Chance: Modeling Density to Enhance Scatter Plot Quality Through Random Data Sampling 20
Data Mining! Computational methods to automatically extract valuable information from raw data! Approaches:! supervised learning uses training samples! examples: decision trees, neural networks! unsupervised learning! example: cluster analysis! association rule mining! dimensionality reduction! Visual data mining! interactive visualization, to help with parameter specification 21 Data Mining Papers (Feb 22-Mar 1)! From Visual Data Exploration To Visual Data Mining: A Survey! Investigating and Reflecting on the Integration of Automatic Data Analysis and Visualization in Knowledge Discovery! Extreme Visualization: Squeezing a Billion Records into a Million Pixels 22
Spatio-Temporal Data Analysis! Spatio-temporal! involves both space and time! Complexities of scale and uncertainty! scale maps to look for patterns and trends! data is often incomplete, collected at different times 23 Spatio-Temporal Data Analysis Papers (Mar 1-22)! Interactive Visual Clustering of Large Collections of Trajectories! Data Mining on Temporal Data: A Visual Approach and its Clinical Application to Hemodialysis! Interactive Pattern Search in Time Series! TimeMatrix: Visualizing Temporal Social Networks Using Interactive Matrix-Based Visualizations 24
Infrastructure / Tools! Linking together all the processes, functions, and services required by visual analytics applications! Require high interactivity! Most visual analytics applications are custombuilt, stand-alone 25 Infrastructure / Tools Papers (Mar 22-29)! Polaris: A System for Query, Analysis, and Visualization of Multidimensional Relational Databases! Jigsaw: Supporting Investigative Analysis Through Interactive Visualization! Collaborative Brushing and Linking for Co- Located Visual Analytics of Document Collections 26
Perception and Cognition! Human side of visual analytics! Visual perception! means by which people interpret their surroundings (or images on a computer screen)! Cognition! ability to understand the visual information and make inferences! Knowledge of how we "think visually" is important in the design of user interfaces 27 Perception and Cognition Papers (Apr 5-12)! Characterizing Users' Visual Analytic Activity for Insight Provenance! Distributed Cognition as a Theoretical Framework for Information Visualization! Toward a Deeper Understanding of the Role of Interaction in Information Visualization 28
Evaluation! What makes a good visual analytics system?! How can we compare systems?! Evaluation is very difficult given! the explorative nature of visual analytics! the wide range of user experience! the diversity of data sources! the actual tasks themselves 29 Evaluation Papers (Apr 12-19)! Evaluating Information Visualization! An Insight-Based Methodology for Evaluating Bioinformatics Visualizations! An Insight-Based Longitudinal Study of Visual Analytics 30
The Future All grand challenge problems of the 21 st century, such as climate change, energy, financial, health, security, require the exploration and analysis of very large and complex data sets which can neither be done by the computer nor the human alone. 31 CS 796/896 Going Forward! Today! research methods webpage! more visualization videos! Next Tuesday Jan 18! Martin Klein Intro to R Tutorial! Demos of some tools, interesting websites! May be meeting in 3 rd floor OR Lab (sign-up for the mailing list and watch your email for details)! Jan 25 First Student Presentations 32
More Video! Hans Rosling's TED 2006 talk! http://www.ted.com/index.php/talks/ hans_rosling_shows_the_best_stats_you_ve_ever_s een.html! David McCandless's TED 2010 talk! http://www.ted.com/talks/ david_mccandless_the_beauty_of_data_visualizatio n.html 33